Compare commits

..

3 Commits

Author SHA1 Message Date
Joseph Doherty e13152f340 test: remove redundant HostBuildingCollection workaround (shared lib no longer installs a global frozen logger) 2026-06-01 08:47:21 -04:00
Joseph Doherty deba5ed115 refactor(logging): correlation scope + redaction on shared ILogRedactor seam
Move the per-request correlation context and secret redaction off the MEL
mechanism onto the Serilog primitives the shared bootstrap consumes.

- GatewayRequestLoggingMiddlewareExtensions now pushes the correlation
  properties (SessionId / WorkerProcessId / CorrelationId / CommandMethod /
  ClientIdentity) via Serilog LogContext.PushProperty for the request lifetime
  and pops them on completion, replacing MEL ILogger.BeginScope. Header parsing
  and property names are unchanged; GatewayLogScope remains the data holder.
- Add GatewayLogRedactorAdapter : ILogRedactor delegating to the existing
  GatewayLogRedactor policy (mxgw_ bearer tokens / credential-bearing command
  values), registered as a singleton so the shared RedactionEnricher masks
  secrets on every event. Remove the now-dead GatewayLoggerExtensions MEL helper.
- Tests: add GatewayLogRedactorAdapterTests; serialize the four host-building
  test classes into one non-parallel collection (HostBuildingCollection) so the
  process-wide Serilog bootstrap logger is not frozen by two concurrent host
  builds racing in parallel collections.

The net48/x86 worker is untouched.
2026-06-01 08:06:28 -04:00
Joseph Doherty 4bf71a0b2c refactor(logging): adopt ZB.MOM.WW.Telemetry.Serilog bootstrap
Swap the gateway process logging from the default Microsoft.Extensions.Logging
provider onto the shared ZB.MOM.WW.Telemetry.Serilog two-stage bootstrap.

- Add a cross-repo ProjectReference to ZB.MOM.WW.Telemetry.Serilog (transitively
  brings the Telemetry core package); the referenced project resolves its own
  Directory.Build.props / Directory.Packages.props so it does not perturb this build.
- Replace MEL wiring in GatewayApplication with builder.AddZbSerilog(ServiceName=
  "mxgateway"; SiteId/NodeRole read from MxGateway:Telemetry when present) and add
  app.UseSerilogRequestLogging().
- Add a Serilog section (Console + daily rolling File sinks, MinimumLevel) to
  appsettings.json and a MinimumLevel override to appsettings.Development.json,
  replacing the old MEL Logging sections.

The net48/x86 worker is untouched. Correlation scope + redaction move to the
shared ILogRedactor seam in the follow-up commit.
2026-06-01 08:03:49 -04:00
56 changed files with 328 additions and 3685 deletions
-19
View File
@@ -107,7 +107,6 @@ public sealed class MxGatewayClientOptions
public required string ApiKey { get; init; }
public bool UseTls { get; init; }
public string? CaCertificatePath { get; init; }
public bool RequireCertificateValidation { get; init; }
public string? ServerNameOverride { get; init; }
public TimeSpan ConnectTimeout { get; init; } = TimeSpan.FromSeconds(10);
public TimeSpan DefaultCallTimeout { get; init; } = TimeSpan.FromSeconds(30);
@@ -125,24 +124,6 @@ or subscription changes because those calls can partially succeed in MXAccess.
API key may be loaded from `MXGATEWAY_API_KEY` by the CLI, not implicitly by the
library constructor unless a helper explicitly says it does that.
### TLS trust posture
The gateway can serve a self-signed certificate it generates itself (it has no
PKI). To make that usable, TLS is **lenient by default**: when `UseTls` is set
and `CaCertificatePath` is empty, `CreateHttpHandler` installs a
`RemoteCertificateValidationCallback` that returns `true`, so the gateway's
self-signed certificate is accepted without verification.
To verify the gateway instead:
- set `CaCertificatePath` to pin a CA — validated via a `CustomRootTrust`
`X509Chain` against that root, and the callback additionally rejects a
hostname/SAN mismatch (`RemoteCertificateNameMismatch`); or
- set `RequireCertificateValidation` to `true` to keep the default OS/system-trust
verification on a connection with no pinned CA.
Pinning a CA always wins over the lenient default.
## Auth Interceptor
Use a gRPC call credentials/interceptor layer to attach:
-11
View File
@@ -287,17 +287,6 @@ Use TLS options for a secured gateway:
dotnet run --project clients/dotnet/ZB.MOM.WW.MxGateway.Client.Cli -- smoke --endpoint https://ZB.MOM.WW.MxGateway.example.local:5001 --tls --ca-file C:\certs\mxgateway-ca.pem --server-name ZB.MOM.WW.MxGateway.example.local --api-key-env MXGATEWAY_API_KEY --item Area001.Pump001.Speed --json
```
### TLS trust
The gateway can auto-generate its own self-signed certificate (it has no PKI), so
the client is **lenient by default**: a TLS connection (`UseTls` / `--tls`) with
no pinned CA accepts whatever certificate the gateway presents. To verify
instead, pin a CA with `CaCertificatePath` / `--ca-file` (this path also enforces
the certificate hostname/SAN match), or set `RequireCertificateValidation` to
force OS/system-trust verification without pinning. Use `ServerNameOverride` /
`--server-name` when the dialed host differs from the certificate SAN. See
[Gateway Configuration](../../docs/GatewayConfiguration.md#automatic-self-signed-certificate).
## Integration Checks
Run live checks only when a gateway and MXAccess-backed worker are available:
@@ -1,85 +0,0 @@
using System.Net.Http;
using System.Net.Security;
using ZB.MOM.WW.MxGateway.Client;
namespace ZB.MOM.WW.MxGateway.Client.Tests;
public sealed class MxGatewayClientTlsHandlerTests
{
/// <summary>
/// Verifies that when TLS is used with no pinned CA and RequireCertificateValidation is false (default),
/// the handler installs an accept-all callback so the gateway's self-signed cert is trusted.
/// The callback must return true regardless of chain errors.
/// </summary>
[Fact]
public void Handler_SkipsVerification_WhenTlsAndNoCaPinned()
{
MxGatewayClientOptions options = new()
{
Endpoint = new Uri("https://localhost:5120"),
ApiKey = "k",
UseTls = true,
};
using SocketsHttpHandler handler = MxGatewayClient.CreateHttpHandlerForTests(options);
Assert.NotNull(handler.SslOptions.RemoteCertificateValidationCallback);
Assert.True(handler.SslOptions.RemoteCertificateValidationCallback!(null!, null!, null, SslPolicyErrors.RemoteCertificateChainErrors));
}
/// <summary>
/// Verifies that when RequireCertificateValidation is true, the callback is left null
/// so the OS trust store performs validation.
/// </summary>
[Fact]
public void Handler_KeepsDefaultVerification_WhenRequireCertificateValidation()
{
MxGatewayClientOptions options = new()
{
Endpoint = new Uri("https://localhost:5120"),
ApiKey = "k",
UseTls = true,
RequireCertificateValidation = true,
};
using SocketsHttpHandler handler = MxGatewayClient.CreateHttpHandlerForTests(options);
Assert.Null(handler.SslOptions.RemoteCertificateValidationCallback);
}
}
public sealed class GalaxyRepositoryClientTlsHandlerTests
{
/// <summary>
/// Verifies that when TLS is used with no pinned CA and RequireCertificateValidation is false (default),
/// the Galaxy client handler installs an accept-all callback so the gateway's self-signed cert is trusted.
/// The callback must return true regardless of chain errors.
/// </summary>
[Fact]
public void Handler_SkipsVerification_WhenTlsAndNoCaPinned()
{
MxGatewayClientOptions options = new()
{
Endpoint = new Uri("https://localhost:5120"),
ApiKey = "k",
UseTls = true,
};
using SocketsHttpHandler handler = GalaxyRepositoryClient.CreateHttpHandlerForTests(options);
Assert.NotNull(handler.SslOptions.RemoteCertificateValidationCallback);
Assert.True(handler.SslOptions.RemoteCertificateValidationCallback!(null!, null!, null, SslPolicyErrors.RemoteCertificateChainErrors));
}
/// <summary>
/// Verifies that when RequireCertificateValidation is true, the Galaxy client callback is left null
/// so the OS trust store performs validation.
/// </summary>
[Fact]
public void Handler_KeepsDefaultVerification_WhenRequireCertificateValidation()
{
MxGatewayClientOptions options = new()
{
Endpoint = new Uri("https://localhost:5120"),
ApiKey = "k",
UseTls = true,
RequireCertificateValidation = true,
};
using SocketsHttpHandler handler = GalaxyRepositoryClient.CreateHttpHandlerForTests(options);
Assert.Null(handler.SslOptions.RemoteCertificateValidationCallback);
}
}
@@ -490,10 +490,7 @@ public sealed class GalaxyRepositoryClient : IAsyncDisposable
.ConfigureAwait(false);
}
private static HttpMessageHandler CreateHttpHandler(MxGatewayClientOptions options) =>
CreateHttpHandlerForTests(options);
internal static SocketsHttpHandler CreateHttpHandlerForTests(MxGatewayClientOptions options)
private static HttpMessageHandler CreateHttpHandler(MxGatewayClientOptions options)
{
SocketsHttpHandler handler = new()
{
@@ -513,11 +510,6 @@ public sealed class GalaxyRepositoryClient : IAsyncDisposable
X509Certificate2 trustedRoot = X509CertificateLoader.LoadCertificateFromFile(options.CaCertificatePath);
handler.SslOptions.RemoteCertificateValidationCallback = (_, certificate, chain, errors) =>
{
if ((errors & System.Net.Security.SslPolicyErrors.RemoteCertificateNameMismatch) != 0)
{
return false;
}
if (certificate is null)
{
return false;
@@ -533,10 +525,6 @@ public sealed class GalaxyRepositoryClient : IAsyncDisposable
return customChain.Build(certificateToValidate);
};
}
else if (!options.RequireCertificateValidation)
{
handler.SslOptions.RemoteCertificateValidationCallback = (_, _, _, _) => true;
}
}
return handler;
@@ -315,10 +315,7 @@ public sealed class MxGatewayClient : IAsyncDisposable
.ConfigureAwait(false);
}
private static HttpMessageHandler CreateHttpHandler(MxGatewayClientOptions options) =>
CreateHttpHandlerForTests(options);
internal static SocketsHttpHandler CreateHttpHandlerForTests(MxGatewayClientOptions options)
private static HttpMessageHandler CreateHttpHandler(MxGatewayClientOptions options)
{
SocketsHttpHandler handler = new()
{
@@ -338,11 +335,6 @@ public sealed class MxGatewayClient : IAsyncDisposable
X509Certificate2 trustedRoot = X509CertificateLoader.LoadCertificateFromFile(options.CaCertificatePath);
handler.SslOptions.RemoteCertificateValidationCallback = (_, certificate, chain, errors) =>
{
if ((errors & System.Net.Security.SslPolicyErrors.RemoteCertificateNameMismatch) != 0)
{
return false;
}
if (certificate is null)
{
return false;
@@ -358,10 +350,6 @@ public sealed class MxGatewayClient : IAsyncDisposable
return customChain.Build(certificateToValidate);
};
}
else if (!options.RequireCertificateValidation)
{
handler.SslOptions.RemoteCertificateValidationCallback = (_, _, _, _) => true;
}
}
return handler;
@@ -27,14 +27,6 @@ public sealed class MxGatewayClientOptions
/// </summary>
public string? CaCertificatePath { get; init; }
/// <summary>
/// When true, TLS connections without a pinned <see cref="CaCertificatePath"/>
/// use the OS trust store. When false (default), the gateway certificate is
/// accepted without verification — appropriate for this internal tool's
/// auto-generated self-signed certificate. Pinning a CA always verifies.
/// </summary>
public bool RequireCertificateValidation { get; init; }
/// <summary>
/// Gets the server name override for SNI during TLS handshake.
/// </summary>
@@ -27,10 +27,4 @@
<None Include="..\README.md" Pack="true" PackagePath="\" />
</ItemGroup>
<ItemGroup>
<AssemblyAttribute Include="System.Runtime.CompilerServices.InternalsVisibleTo">
<_Parameter1>ZB.MOM.WW.MxGateway.Client.Tests</_Parameter1>
</AssemblyAttribute>
</ItemGroup>
</Project>
-17
View File
@@ -104,23 +104,6 @@ Support:
- `credentials.NewClientTLSFromFile`,
- custom `tls.Config` for advanced callers.
### Trust posture
The gateway can serve a self-signed certificate it generates itself (it has no
PKI). To make that usable, TLS is **lenient by default**: when `Plaintext` is
`false` and no `CACertFile`/`TLSConfig`/`TransportCredentials` is supplied,
`buildCredentials` dials with `tls.Config{InsecureSkipVerify: true}` (carrying
`ServerNameOverride` as the SNI when set), so the gateway's self-signed
certificate is accepted without verification.
To verify the gateway instead:
- set `CACertFile` to pin a CA (full verification against that root), or
- set `RequireCertificateValidation: true` to verify against the OS/system trust
roots without pinning.
Pinning a CA always wins over the lenient default.
## Streaming
`Events(ctx)` should return a receive channel of:
-8
View File
@@ -75,14 +75,6 @@ client, err := mxgateway.Dial(ctx, mxgateway.Options{
})
```
The gateway can auto-generate its own self-signed certificate (it has no PKI), so
the client is **lenient by default**: a TLS connection (`Plaintext: false`) with
no `CACertFile`/`TLSConfig` accepts whatever certificate the gateway presents
(`InsecureSkipVerify`, with `ServerNameOverride` as the SNI when set). To verify
instead, set `CACertFile` to pin a CA, or set `RequireCertificateValidation:
true` to verify against the OS/system trust roots without pinning. See
[Gateway Configuration](../../docs/GatewayConfiguration.md#automatic-self-signed-certificate).
`Client.OpenSession` returns a `Session` with helpers for `Register`,
`AddItem`, `AddItem2`, `Advise`, `Write`, `Events`, and `Close`. Prefer
`SubscribeEvents` or `SubscribeEventsAfter` for long-running streams because the
+4 -16
View File
@@ -222,22 +222,10 @@ func resolveTransportCredentials(opts Options) (credentials.TransportCredentials
return credentials.NewTLS(cfg), nil
}
return credentials.NewTLS(tlsConfigForOptions(opts)), nil
}
// tlsConfigForOptions returns the *tls.Config for the no-CA, no-custom-config TLS path.
// It returns nil when the caller should use a different credentials path (CA file or custom TLSConfig).
// Exposed as an internal helper so unit tests can assert the InsecureSkipVerify posture.
func tlsConfigForOptions(opts Options) *tls.Config {
// CA file and custom TLSConfig take their own paths in resolveTransportCredentials.
if opts.CACertFile != "" || opts.TLSConfig != nil {
return nil
}
return &tls.Config{
MinVersion: tls.VersionTLS12,
ServerName: opts.ServerNameOverride,
InsecureSkipVerify: !opts.RequireCertificateValidation, //nolint:gosec // internal tool; self-signed gateway cert expected; opt-in strict via RequireCertificateValidation
}
return credentials.NewTLS(&tls.Config{
MinVersion: tls.VersionTLS12,
ServerName: opts.ServerNameOverride,
}), nil
}
// OpenSessionOptions describes fields used to create an OpenSessionRequest.
-59
View File
@@ -1,59 +0,0 @@
package mxgateway
import (
"crypto/tls"
"testing"
)
// tlsConfigFromOptions is the internal helper under test.
// It extracts the *tls.Config from the no-CA TLS path of resolveTransportCredentials.
// We exercise it directly to avoid needing a real dial target.
func TestTLSInsecureSkipVerify_DefaultTrue(t *testing.T) {
cfg := tlsConfigForOptions(Options{
Endpoint: "localhost:5120",
})
if cfg == nil {
t.Fatal("expected non-nil tls.Config")
}
if !cfg.InsecureSkipVerify {
t.Error("InsecureSkipVerify should be true by default when no CA is pinned")
}
}
func TestTLSInsecureSkipVerify_FalseWhenRequireCertificateValidation(t *testing.T) {
cfg := tlsConfigForOptions(Options{
Endpoint: "localhost:5120",
RequireCertificateValidation: true,
})
if cfg == nil {
t.Fatal("expected non-nil tls.Config")
}
if cfg.InsecureSkipVerify {
t.Error("InsecureSkipVerify should be false when RequireCertificateValidation is true")
}
}
func TestTLSInsecureSkipVerify_FalseWhenCACertFileSet(t *testing.T) {
// When a CA file is pinned, the CA-verification path is taken instead.
// tlsConfigForOptions should return nil (the CA path does not use our helper).
cfg := tlsConfigForOptions(Options{
Endpoint: "localhost:5120",
CACertFile: "/some/ca.pem",
})
if cfg != nil {
t.Error("expected nil tls.Config when CACertFile is set (CA path taken)")
}
}
func TestTLSInsecureSkipVerify_FalseWhenCustomTLSConfig(t *testing.T) {
// When TLSConfig is supplied explicitly, our default skip-verify must not overwrite it.
custom := &tls.Config{MinVersion: tls.VersionTLS13}
cfg := tlsConfigForOptions(Options{
Endpoint: "localhost:5120",
TLSConfig: custom,
})
if cfg != nil {
t.Error("expected nil tls.Config when TLSConfig is already set (custom config path taken)")
}
}
-4
View File
@@ -34,10 +34,6 @@ type Options struct {
TransportCredentials credentials.TransportCredentials
// DialOptions are appended to the gRPC dial options after the defaults.
DialOptions []grpc.DialOption
// RequireCertificateValidation forces TLS certificate verification even when
// no CACertFile is pinned. Default false: the gateway's self-signed cert is
// accepted without verification (internal-tool posture).
RequireCertificateValidation bool
}
// BrowseChildrenOptions configures lazy Galaxy hierarchy walks performed by
-17
View File
@@ -112,23 +112,6 @@ Support:
- custom CA certificate file,
- server name override for test environments.
### Trust posture
The gateway can serve a self-signed certificate it generates itself (it has no
PKI). To make that usable, TLS is **lenient by default**: when the channel is not
plaintext and no `caCertificatePath` is set, the client builds
`GrpcSslContexts.forClient().trustManager(InsecureTrustManagerFactory.INSTANCE)`
(grpc-netty-shaded), so the gateway's self-signed certificate is accepted without
verification.
To verify the gateway instead:
- set `caCertificatePath` to pin a CA (full verification against that root), or
- set `requireCertificateValidation` to `true` to verify against the JVM trust
store without pinning.
Pinning a CA always wins over the lenient default.
## Streaming
Support both:
-10
View File
@@ -57,16 +57,6 @@ try (MxGatewayClient client = MxGatewayClient.connect(options);
}
```
The gateway can auto-generate its own self-signed certificate (it has no PKI), so
the client is **lenient by default**: a TLS connection (`plaintext(false)`) with
no `caCertificatePath` accepts whatever certificate the gateway presents (via
grpc-netty-shaded's `InsecureTrustManagerFactory`). To verify instead, set
`caCertificatePath` to pin a CA, or set `requireCertificateValidation(true)` to
verify against the JVM trust store without pinning. Use `serverNameOverride` /
`--server-name-override` when the dialed host differs from the certificate SAN.
See
[Gateway Configuration](../../docs/GatewayConfiguration.md#automatic-self-signed-certificate).
Use `rawBlockingStub`, `rawFutureStub`, `rawAsyncStub`, `openSessionRaw`,
`closeSessionRaw`, `invoke`, and raw session helper methods when tests need the
underlying protobuf messages. `MxGatewayCommandException` and
-4
View File
@@ -9,10 +9,6 @@ pluginManagement {
}
}
plugins {
id 'org.gradle.toolchains.foojay-resolver-convention' version '1.0.0'
}
dependencyResolutionManagement {
repositoriesMode.set(RepositoriesMode.FAIL_ON_PROJECT_REPOS)
repositories {
@@ -384,15 +384,6 @@ public final class MxGatewayClient implements AutoCloseable {
} catch (SSLException error) {
throw new MxGatewayException("failed to configure gateway TLS", error);
}
} else if (!options.requireCertificateValidation()) {
try {
builder.sslContext(GrpcSslContexts.forClient()
.trustManager(io.grpc.netty.shaded.io.netty.handler.ssl.util
.InsecureTrustManagerFactory.INSTANCE)
.build());
} catch (SSLException error) {
throw new MxGatewayException("failed to configure lenient gateway TLS", error);
}
} else {
builder.useTransportSecurity();
}
@@ -402,19 +393,6 @@ public final class MxGatewayClient implements AutoCloseable {
return builder.build();
}
/**
* Package-visible test seam — creates a raw {@link ManagedChannel} from the
* given options without attaching auth interceptors. Used by TLS fixture
* tests to verify channel construction behaviour without a full
* {@link MxGatewayClient} wrapper.
*
* @param options the client options
* @return a new {@link ManagedChannel}
*/
static ManagedChannel createChannelForTests(MxGatewayClientOptions options) {
return createChannel(options);
}
private <T extends io.grpc.stub.AbstractStub<T>> T withDeadline(T stub) {
if (options.callTimeout().isNegative()) {
return stub;
@@ -20,7 +20,6 @@ public final class MxGatewayClientOptions {
private final String apiKey;
private final boolean plaintext;
private final Path caCertificatePath;
private final boolean requireCertificateValidation;
private final String serverNameOverride;
private final Duration connectTimeout;
private final Duration callTimeout;
@@ -32,7 +31,6 @@ public final class MxGatewayClientOptions {
apiKey = builder.apiKey == null ? "" : builder.apiKey;
plaintext = builder.plaintext;
caCertificatePath = builder.caCertificatePath;
requireCertificateValidation = builder.requireCertificateValidation;
serverNameOverride = builder.serverNameOverride == null ? "" : builder.serverNameOverride;
connectTimeout = builder.connectTimeout == null ? DEFAULT_CONNECT_TIMEOUT : builder.connectTimeout;
callTimeout = builder.callTimeout == null ? DEFAULT_CALL_TIMEOUT : builder.callTimeout;
@@ -97,18 +95,6 @@ public final class MxGatewayClientOptions {
return caCertificatePath;
}
/**
* Returns whether TLS certificate verification is required even when no CA is pinned.
* When {@code false} (default), the gateway's self-signed certificate is accepted
* without verification. When {@code true}, the OS trust store is used.
* Pinning a CA via {@link #caCertificatePath()} always verifies regardless of this flag.
*
* @return {@code true} if strict certificate verification is required
*/
public boolean requireCertificateValidation() {
return requireCertificateValidation;
}
/**
* Returns the TLS server-name override, or an empty string when none was supplied.
*
@@ -162,8 +148,6 @@ public final class MxGatewayClientOptions {
+ plaintext
+ ", caCertificatePath="
+ caCertificatePath
+ ", requireCertificateValidation="
+ requireCertificateValidation
+ ", serverNameOverride='"
+ serverNameOverride
+ '\''
@@ -193,7 +177,6 @@ public final class MxGatewayClientOptions {
private String apiKey;
private boolean plaintext;
private Path caCertificatePath;
private boolean requireCertificateValidation;
private String serverNameOverride;
private Duration connectTimeout;
private Duration callTimeout;
@@ -247,21 +230,6 @@ public final class MxGatewayClientOptions {
return this;
}
/**
* When {@code true}, TLS connections without a pinned CA use the OS trust store
* and will reject the gateway's self-signed certificate. When {@code false}
* (default), the gateway certificate is accepted without verification —
* appropriate for this internal tool's auto-generated self-signed certificate.
* Pinning a CA via {@link #caCertificatePath(Path)} always verifies.
*
* @param value {@code true} to require certificate validation, {@code false} to accept any cert
* @return this builder
*/
public Builder requireCertificateValidation(boolean value) {
requireCertificateValidation = value;
return this;
}
/**
* Overrides the TLS server name used during the handshake.
*
@@ -1,198 +0,0 @@
package com.zb.mom.ww.mxgateway.client;
import static org.junit.jupiter.api.Assertions.assertThrows;
import static org.junit.jupiter.api.Assertions.assertTrue;
import io.grpc.ManagedChannel;
import io.grpc.Server;
import io.grpc.StatusRuntimeException;
import io.grpc.netty.shaded.io.grpc.netty.GrpcSslContexts;
import io.grpc.netty.shaded.io.grpc.netty.NettyServerBuilder;
import io.grpc.stub.StreamObserver;
import java.io.File;
import java.io.FileOutputStream;
import java.io.IOException;
import java.net.InetSocketAddress;
import java.nio.file.Files;
import java.security.KeyStore;
import java.security.PrivateKey;
import java.security.cert.Certificate;
import java.security.cert.X509Certificate;
import java.time.Duration;
import java.util.Base64;
import java.util.concurrent.TimeUnit;
import javax.net.ssl.SSLException;
import mxaccess_gateway.v1.MxAccessGatewayGrpc;
import mxaccess_gateway.v1.MxaccessGateway.OpenSessionReply;
import mxaccess_gateway.v1.MxaccessGateway.OpenSessionRequest;
import mxaccess_gateway.v1.MxaccessGateway.ProtocolStatus;
import mxaccess_gateway.v1.MxaccessGateway.ProtocolStatusCode;
import org.junit.jupiter.api.AfterEach;
import org.junit.jupiter.api.BeforeEach;
import org.junit.jupiter.api.Test;
/**
* Verifies that the Java client connects to a Netty TLS server with a
* self-signed certificate when no CA is pinned (lenient default), and that
* setting {@code requireCertificateValidation(true)} causes a TLS failure.
*
* <p>A self-signed certificate is generated using {@code keytool} (always
* available in the JDK) to avoid dependencies on internal JDK APIs or
* BouncyCastle, and so the test works on all JDK versions used by the project.
*/
final class MxGatewayClientTlsTests {
private Server server;
private int port;
private File certPemFile;
private File keyPemFile;
private File keystoreFile;
@BeforeEach
void startTlsServer() throws Exception {
keystoreFile = File.createTempFile("gw-test-ks", ".p12");
certPemFile = File.createTempFile("gw-test-cert", ".pem");
keyPemFile = File.createTempFile("gw-test-key", ".pem");
// keytool refuses to write to a pre-existing (even empty) file; delete it first.
keystoreFile.delete();
// Use keytool to generate a self-signed PKCS12 keystore.
String keytool = ProcessHandle.current().info().command()
.map(cmd -> cmd.replace("java", "keytool"))
.orElse("keytool");
// Fall back to just "keytool" on PATH if the resolved path doesn't exist.
if (!new File(keytool).exists()) {
keytool = "keytool";
}
Process p = new ProcessBuilder(
keytool,
"-genkeypair",
"-alias", "server",
"-keyalg", "RSA",
"-keysize", "2048",
"-sigalg", "SHA256withRSA",
"-validity", "1",
"-dname", "CN=localhost",
"-storetype", "PKCS12",
"-storepass", "changeit",
"-keypass", "changeit",
"-keystore", keystoreFile.getAbsolutePath())
.redirectErrorStream(true)
.start();
int exit = p.waitFor();
if (exit != 0) {
String out = new String(p.getInputStream().readAllBytes());
throw new IllegalStateException("keytool failed (exit " + exit + "): " + out);
}
// Export cert and private key from the PKCS12 keystore to PEM files.
KeyStore ks = KeyStore.getInstance("PKCS12");
try (var is = Files.newInputStream(keystoreFile.toPath())) {
ks.load(is, "changeit".toCharArray());
}
X509Certificate cert = (X509Certificate) ks.getCertificate("server");
PrivateKey privateKey = (PrivateKey) ks.getKey("server", "changeit".toCharArray());
try (FileOutputStream out = new FileOutputStream(certPemFile)) {
out.write("-----BEGIN CERTIFICATE-----\n".getBytes());
out.write(Base64.getMimeEncoder(64, new byte[]{'\n'}).encode(cert.getEncoded()));
out.write("\n-----END CERTIFICATE-----\n".getBytes());
}
try (FileOutputStream out = new FileOutputStream(keyPemFile)) {
out.write("-----BEGIN PRIVATE KEY-----\n".getBytes());
out.write(Base64.getMimeEncoder(64, new byte[]{'\n'}).encode(privateKey.getEncoded()));
out.write("\n-----END PRIVATE KEY-----\n".getBytes());
}
server = NettyServerBuilder
.forAddress(new InetSocketAddress("127.0.0.1", 0))
.sslContext(GrpcSslContexts.forServer(certPemFile, keyPemFile).build())
.addService(new MinimalGatewayService())
.build()
.start();
port = server.getPort();
}
@AfterEach
void stopTlsServer() throws InterruptedException {
if (server != null) {
server.shutdown();
server.awaitTermination(5, TimeUnit.SECONDS);
}
if (certPemFile != null) {
certPemFile.delete();
}
if (keyPemFile != null) {
keyPemFile.delete();
}
if (keystoreFile != null) {
keystoreFile.delete();
}
}
@Test
void connectsToSelfSignedServer_WhenRequireCertificateValidationIsFalse() throws SSLException {
// Default options — requireCertificateValidation defaults to false.
MxGatewayClientOptions options = MxGatewayClientOptions.builder()
.endpoint("127.0.0.1:" + port)
.apiKey("test-key")
.connectTimeout(Duration.ofSeconds(5))
.callTimeout(Duration.ofSeconds(5))
.build();
ManagedChannel channel = MxGatewayClient.createChannelForTests(options);
try {
MxAccessGatewayGrpc.MxAccessGatewayBlockingStub stub =
MxAccessGatewayGrpc.newBlockingStub(channel);
OpenSessionReply reply = stub.openSession(
OpenSessionRequest.newBuilder()
.setClientSessionName("tls-test")
.build());
assertTrue(reply.getProtocolStatus().getCode()
== ProtocolStatusCode.PROTOCOL_STATUS_CODE_OK);
} finally {
channel.shutdownNow();
}
}
@Test
void failsToConnect_WhenRequireCertificateValidationIsTrue() throws SSLException {
MxGatewayClientOptions options = MxGatewayClientOptions.builder()
.endpoint("127.0.0.1:" + port)
.apiKey("test-key")
.requireCertificateValidation(true)
.connectTimeout(Duration.ofSeconds(5))
.callTimeout(Duration.ofSeconds(5))
.build();
ManagedChannel channel = MxGatewayClient.createChannelForTests(options);
try {
MxAccessGatewayGrpc.MxAccessGatewayBlockingStub stub =
MxAccessGatewayGrpc.newBlockingStub(channel);
assertThrows(StatusRuntimeException.class, () ->
stub.openSession(OpenSessionRequest.newBuilder()
.setClientSessionName("tls-strict-test")
.build()));
} finally {
channel.shutdownNow();
}
}
/** Minimal gateway stub that succeeds any OpenSession call. */
private static final class MinimalGatewayService
extends MxAccessGatewayGrpc.MxAccessGatewayImplBase {
@Override
public void openSession(
OpenSessionRequest request,
StreamObserver<OpenSessionReply> responseObserver) {
responseObserver.onNext(OpenSessionReply.newBuilder()
.setSessionId("tls-test-session")
.setProtocolStatus(ProtocolStatus.newBuilder()
.setCode(ProtocolStatusCode.PROTOCOL_STATUS_CODE_OK)
.build())
.build());
responseObserver.onCompleted();
}
}
}
-22
View File
@@ -112,28 +112,6 @@ Support:
- TLS channel with default roots,
- custom root certificate file.
### Trust posture (trust-on-first-use)
The gateway can serve a self-signed certificate it generates itself (it has no
PKI). grpc-python exposes no per-channel skip-verify hook, so the client cannot
"accept any certificate" the way the other clients do. Instead, when the channel
is not plaintext and neither `ca_file` nor `require_certificate_validation` is
set, the TLS default is **trust-on-first-use**: the client fetches the server's
presented certificate once via `ssl.get_server_certificate` (an unverified
probe), pins it as the channel's only trust root, and — because the generated
certificate always carries a `localhost` SAN — defaults
`grpc.ssl_target_name_override` to `localhost` when no `server_name_override` was
supplied (tolerating dial-by-IP or a hostname mismatch). A failed probe is
surfaced as a transport error naming the endpoint.
To verify the gateway instead:
- set `ca_file` to verify against a specific CA, or
- set `require_certificate_validation=True` to verify against the system trust
roots.
Both bypass the TOFU path.
## Streaming
Expose `stream_events` as an async iterator. Canceling the task should cancel
-11
View File
@@ -230,17 +230,6 @@ The client supports plaintext channels for local development, TLS with system
roots, TLS with a custom `ca_file`, and an optional test server name override.
API keys are redacted from option repr output and CLI error output.
The gateway can auto-generate its own self-signed certificate (it has no PKI).
grpc-python has no per-channel skip-verify, so the lenient TLS default is
**trust-on-first-use**: with no `ca_file` and `require_certificate_validation`
left `False`, the client fetches the gateway's presented certificate once
(unverified) and pins it for the channel, defaulting the SNI/target-name override
to `localhost` (the generated certificate always carries a `localhost` SAN) when
none was supplied. To verify instead, pass `ca_file` to verify against a specific
CA, or set `require_certificate_validation=True` to verify against the system
trust roots. See
[Gateway Configuration](../../docs/GatewayConfiguration.md#automatic-self-signed-certificate).
## CLI
The CLI emits deterministic JSON for automation:
@@ -2,7 +2,6 @@
from __future__ import annotations
import ssl
from collections.abc import Sequence
from dataclasses import dataclass, field
from pathlib import Path
@@ -10,7 +9,6 @@ from pathlib import Path
import grpc
from .auth import REDACTED, ApiKey
from .errors import MxGatewayTransportError
@dataclass(frozen=True)
@@ -21,7 +19,6 @@ class ClientOptions:
api_key: str | ApiKey | None = None
plaintext: bool = False
ca_file: str | None = None
require_certificate_validation: bool = False
server_name_override: str | None = None
call_timeout: float | None = 30.0
stream_timeout: float | None = None
@@ -48,7 +45,6 @@ class ClientOptions:
f"{type(self).__name__}(endpoint={self.endpoint!r}, "
f"api_key={api_key!r}, plaintext={self.plaintext!r}, "
f"ca_file={self.ca_file!r}, "
f"require_certificate_validation={self.require_certificate_validation!r}, "
f"server_name_override={self.server_name_override!r}, "
f"call_timeout={self.call_timeout!r}, "
f"stream_timeout={self.stream_timeout!r}, "
@@ -73,34 +69,8 @@ class BrowseChildrenOptions:
historized_only: bool = False
def _split_authority(endpoint: str) -> tuple[str, int]:
"""Split a gRPC target (optionally scheme-prefixed) into (host, port).
Handles bracketed IPv6 literals (e.g. ``[::1]:5120`` or bare ``[::1]``),
returning the host without brackets so it is safe to pass to
``ssl.get_server_certificate``.
"""
target = endpoint.split("://", 1)[-1]
if target.startswith("["):
# Bracketed IPv6: "[::1]:5120" or "[::1]"
bracket_end = target.find("]")
host = target[1:bracket_end] # strip surrounding brackets
remainder = target[bracket_end + 1 :] # ":5120" or ""
port_str = remainder.lstrip(":")
return (host, int(port_str) if port_str else 443)
host, _, port = target.rpartition(":")
return (host or "localhost", int(port) if port else 443)
def create_channel(options: ClientOptions) -> grpc.aio.Channel:
"""Create a plaintext or TLS `grpc.aio` channel from client options.
The TLS default is lenient: grpc-python has no per-channel skip-verify, so
the server's presented certificate is fetched once (unverified) and pinned
as the channel's only trust root (trust-on-first-use). Set
`require_certificate_validation=True` to force system-trust verification, or
pass `ca_file` to verify against a specific CA — both bypass the TOFU path.
"""
"""Create a plaintext or TLS `grpc.aio` channel from client options."""
channel_options: list[tuple[str, str | int]] = [
("grpc.max_receive_message_length", options.max_grpc_message_bytes),
@@ -112,28 +82,11 @@ def create_channel(options: ClientOptions) -> grpc.aio.Channel:
if options.plaintext:
return grpc.aio.insecure_channel(options.endpoint, options=channel_options)
root_certificates = None
if options.ca_file:
root_certificates = Path(options.ca_file).read_bytes()
credentials = grpc.ssl_channel_credentials(root_certificates=root_certificates)
elif options.require_certificate_validation:
credentials = grpc.ssl_channel_credentials()
else:
# Lenient default: grpc-python has no per-channel skip-verify, so fetch the
# server's certificate (unverified) and pin it for this channel (TOFU).
host, port = _split_authority(options.endpoint)
try:
presented = ssl.get_server_certificate((host, port))
except OSError as error:
raise MxGatewayTransportError(
f"failed to fetch TLS certificate from {options.endpoint}: {error}"
) from error
credentials = grpc.ssl_channel_credentials(root_certificates=presented.encode("ascii"))
# The gateway self-signed cert always carries a "localhost" SAN, so default
# the SNI/target-name override to it when none was supplied, tolerating
# dial-by-IP or hostname mismatch.
if not options.server_name_override:
channel_options.append(("grpc.ssl_target_name_override", "localhost"))
credentials = grpc.ssl_channel_credentials(root_certificates=root_certificates)
return grpc.aio.secure_channel(
options.endpoint,
credentials,
+24 -187
View File
@@ -72,83 +72,27 @@ def test_create_channel_uses_plaintext_channel(monkeypatch: pytest.MonkeyPatch)
]
def test_create_channel_uses_tls_channel_tofu_default(monkeypatch: pytest.MonkeyPatch) -> None:
"""Default TLS (no ca_file, no require_certificate_validation) uses TOFU:
fetches the server cert unverified, pins it as root_certificates, and adds
grpc.ssl_target_name_override = "localhost" automatically.
"""
_DUMMY_PEM = "-----BEGIN CERTIFICATE-----\nZmFrZQ==\n-----END CERTIFICATE-----\n"
get_cert_calls: list[tuple[str, int]] = []
def test_create_channel_uses_tls_channel(monkeypatch: pytest.MonkeyPatch) -> None:
calls: list[tuple[str, object, object]] = []
def fake_get_server_certificate(addr: tuple[str, int]) -> str:
get_cert_calls.append(addr)
return _DUMMY_PEM
cred_calls: list[object] = []
def fake_credentials(*, root_certificates: object = None) -> str:
cred_calls.append(root_certificates)
def fake_credentials(*, root_certificates: object) -> str:
assert root_certificates is None
return "creds"
channel_calls: list[tuple[str, object, object]] = []
def fake_secure_channel(endpoint: str, credentials: object, *, options: object) -> str:
channel_calls.append((endpoint, credentials, options))
calls.append((endpoint, credentials, options))
return "tls-channel"
monkeypatch.setattr(options_module.ssl, "get_server_certificate", fake_get_server_certificate)
monkeypatch.setattr(options_module.grpc, "ssl_channel_credentials", fake_credentials)
monkeypatch.setattr(options_module.grpc.aio, "secure_channel", fake_secure_channel)
channel = create_channel(
ClientOptions(endpoint="gateway.example:5001"),
)
assert channel == "tls-channel"
# TOFU: should have fetched the cert from the server (host, port)
assert get_cert_calls == [("gateway.example", 5001)]
# Pinned the fetched PEM bytes as root_certificates
assert cred_calls == [_DUMMY_PEM.encode("ascii")]
# Auto-injected localhost override (no server_name_override supplied)
assert channel_calls == [
(
"gateway.example:5001",
"creds",
[
("grpc.max_receive_message_length", 16 * 1024 * 1024),
("grpc.max_send_message_length", 16 * 1024 * 1024),
("grpc.ssl_target_name_override", "localhost"),
],
),
]
def test_create_channel_uses_tls_channel_tofu_respects_server_name_override(
monkeypatch: pytest.MonkeyPatch,
) -> None:
"""When server_name_override is set, TOFU still runs but does NOT add the
auto-localhost override (the explicit override is already in channel_options).
"""
_DUMMY_PEM = "-----BEGIN CERTIFICATE-----\nZmFrZQ==\n-----END CERTIFICATE-----\n"
monkeypatch.setattr(
options_module.ssl,
"get_server_certificate",
lambda addr: _DUMMY_PEM,
options_module.grpc,
"ssl_channel_credentials",
fake_credentials,
)
monkeypatch.setattr(
options_module.grpc.aio,
"secure_channel",
fake_secure_channel,
)
cred_calls: list[object] = []
def fake_credentials(*, root_certificates: object = None) -> str:
cred_calls.append(root_certificates)
return "creds"
channel_calls: list[tuple[str, object, object]] = []
def fake_secure_channel(endpoint: str, credentials: object, *, options: object) -> str:
channel_calls.append((endpoint, credentials, options))
return "tls-channel"
monkeypatch.setattr(options_module.grpc, "ssl_channel_credentials", fake_credentials)
monkeypatch.setattr(options_module.grpc.aio, "secure_channel", fake_secure_channel)
channel = create_channel(
ClientOptions(
@@ -158,121 +102,14 @@ def test_create_channel_uses_tls_channel_tofu_respects_server_name_override(
)
assert channel == "tls-channel"
assert cred_calls == [_DUMMY_PEM.encode("ascii")]
assert channel_calls == [
(
"gateway.example:5001",
"creds",
[
("grpc.max_receive_message_length", 16 * 1024 * 1024),
("grpc.max_send_message_length", 16 * 1024 * 1024),
# Explicit override from ClientOptions — not the auto-localhost one
("grpc.ssl_target_name_override", "gateway.test"),
],
),
]
def test_create_channel_uses_tls_channel_require_cert_validation(
monkeypatch: pytest.MonkeyPatch,
) -> None:
"""require_certificate_validation=True uses system trust (no TOFU, no root_certificates)."""
get_cert_called = False
def fake_get_server_certificate(addr: object) -> str: # pragma: no cover
nonlocal get_cert_called
get_cert_called = True
return "SHOULD_NOT_BE_CALLED"
cred_calls: list[object] = []
def fake_credentials(**kwargs: object) -> str:
cred_calls.append(kwargs)
return "creds"
channel_calls: list[tuple[str, object, object]] = []
def fake_secure_channel(endpoint: str, credentials: object, *, options: object) -> str:
channel_calls.append((endpoint, credentials, options))
return "tls-channel"
monkeypatch.setattr(options_module.ssl, "get_server_certificate", fake_get_server_certificate)
monkeypatch.setattr(options_module.grpc, "ssl_channel_credentials", fake_credentials)
monkeypatch.setattr(options_module.grpc.aio, "secure_channel", fake_secure_channel)
channel = create_channel(
ClientOptions(
endpoint="gateway.example:5001",
require_certificate_validation=True,
),
)
assert channel == "tls-channel"
# Must NOT call TOFU prefetch
assert not get_cert_called
# ssl_channel_credentials() called with NO keyword args (system trust)
assert cred_calls == [{}]
assert channel_calls == [
(
"gateway.example:5001",
"creds",
[
("grpc.max_receive_message_length", 16 * 1024 * 1024),
("grpc.max_send_message_length", 16 * 1024 * 1024),
],
),
]
def test_create_channel_uses_tls_channel_ca_file(
monkeypatch: pytest.MonkeyPatch,
tmp_path: pytest.TempPathFactory,
) -> None:
"""ca_file path: reads the PEM file, passes bytes as root_certificates, skips TOFU."""
ca_pem = b"-----BEGIN CERTIFICATE-----\nY2FkYXRh\n-----END CERTIFICATE-----\n"
ca_file = tmp_path / "ca.pem"
ca_file.write_bytes(ca_pem)
get_cert_called = False
def fake_get_server_certificate(addr: object) -> str: # pragma: no cover
nonlocal get_cert_called
get_cert_called = True
return "SHOULD_NOT_BE_CALLED"
cred_calls: list[object] = []
def fake_credentials(*, root_certificates: object = None) -> str:
cred_calls.append(root_certificates)
return "creds"
channel_calls: list[tuple[str, object, object]] = []
def fake_secure_channel(endpoint: str, credentials: object, *, options: object) -> str:
channel_calls.append((endpoint, credentials, options))
return "tls-channel"
monkeypatch.setattr(options_module.ssl, "get_server_certificate", fake_get_server_certificate)
monkeypatch.setattr(options_module.grpc, "ssl_channel_credentials", fake_credentials)
monkeypatch.setattr(options_module.grpc.aio, "secure_channel", fake_secure_channel)
channel = create_channel(
ClientOptions(
endpoint="gateway.example:5001",
ca_file=str(ca_file),
),
)
assert channel == "tls-channel"
assert not get_cert_called
assert cred_calls == [ca_pem]
assert channel_calls == [
(
"gateway.example:5001",
"creds",
[
("grpc.max_receive_message_length", 16 * 1024 * 1024),
("grpc.max_send_message_length", 16 * 1024 * 1024),
],
),
]
assert calls == [
(
"gateway.example:5001",
"creds",
[
("grpc.max_receive_message_length", 16 * 1024 * 1024),
("grpc.max_send_message_length", 16 * 1024 * 1024),
("grpc.ssl_target_name_override", "gateway.test"),
],
),
]
-165
View File
@@ -1,165 +0,0 @@
"""TLS behaviour tests for ``create_channel``.
These spin up a real loopback ``grpc.aio`` server with a freshly generated
self-signed certificate (carrying a ``localhost`` SAN, mirroring the gateway's
auto-generated cert) and assert the lenient TOFU default lets a client connect
without any CA configured.
Marked ``tls`` and skipped unless ``MXGATEWAY_RUN_TLS_TESTS=1`` because loopback
TLS handshakes can be timing-flaky on shared CI runners. This mirrors how the
suite gates anything that depends on real sockets rather than fakes.
"""
from __future__ import annotations
import os
import shutil
import socket
import ssl
import subprocess
import tempfile
from collections.abc import AsyncIterator
from pathlib import Path
import grpc
import pytest
import pytest_asyncio
from zb_mom_ww_mxgateway import ClientOptions
from zb_mom_ww_mxgateway.errors import MxGatewayTransportError
from zb_mom_ww_mxgateway.generated import mxaccess_gateway_pb2 as pb
from zb_mom_ww_mxgateway.generated import mxaccess_gateway_pb2_grpc as pb_grpc
from zb_mom_ww_mxgateway.options import create_channel
pytestmark = pytest.mark.tls
_RUN_TLS_TESTS = os.environ.get("MXGATEWAY_RUN_TLS_TESTS") == "1"
_OPENSSL = shutil.which("openssl")
requires_tls = pytest.mark.skipif(
not _RUN_TLS_TESTS,
reason="set MXGATEWAY_RUN_TLS_TESTS=1 to run loopback TLS tests",
)
requires_openssl = pytest.mark.skipif(
_OPENSSL is None,
reason="openssl CLI is required to generate a self-signed test certificate",
)
def _generate_self_signed_cert(directory: Path) -> tuple[Path, Path]:
"""Generate a self-signed cert/key pair with a ``localhost`` SAN."""
key_path = directory / "server.key"
cert_path = directory / "server.crt"
subprocess.run(
[
str(_OPENSSL),
"req",
"-x509",
"-newkey",
"rsa:2048",
"-nodes",
"-keyout",
str(key_path),
"-out",
str(cert_path),
"-days",
"1",
"-subj",
"/CN=mxgateway-test",
"-addext",
"subjectAltName=DNS:localhost,IP:127.0.0.1",
],
check=True,
capture_output=True,
)
return cert_path, key_path
def _free_port() -> int:
with socket.socket(socket.AF_INET, socket.SOCK_STREAM) as sock:
sock.bind(("127.0.0.1", 0))
return int(sock.getsockname()[1])
class _StaticGatewayServicer(pb_grpc.MxAccessGatewayServicer):
"""Minimal servicer answering ``OpenSession`` with a fixed session id."""
async def OpenSession( # noqa: N802 - generated gRPC method name
self, request: pb.OpenSessionRequest, context: object
) -> pb.OpenSessionReply:
return pb.OpenSessionReply(session_id="tls-session-1")
@pytest_asyncio.fixture
async def tls_server() -> AsyncIterator[int]:
with tempfile.TemporaryDirectory() as tmp:
cert_path, key_path = _generate_self_signed_cert(Path(tmp))
credentials = grpc.ssl_server_credentials(
[(key_path.read_bytes(), cert_path.read_bytes())]
)
server = grpc.aio.server()
pb_grpc.add_MxAccessGatewayServicer_to_server(_StaticGatewayServicer(), server)
port = _free_port()
server.add_secure_port(f"127.0.0.1:{port}", credentials)
await server.start()
try:
yield port
finally:
await server.stop(grace=None)
@requires_tls
@requires_openssl
@pytest.mark.asyncio
async def test_default_tls_connects_via_tofu(tls_server: int) -> None:
"""Default TLS options (no CA) connect by pinning the presented cert."""
options = ClientOptions(
endpoint=f"127.0.0.1:{tls_server}",
api_key="mxgw_test_secret",
)
channel = create_channel(options)
try:
stub = pb_grpc.MxAccessGatewayStub(channel)
reply = await stub.OpenSession(pb.OpenSessionRequest(), timeout=10)
assert reply.session_id == "tls-session-1"
finally:
await channel.close()
def test_split_authority_parses_host_and_port() -> None:
from zb_mom_ww_mxgateway.options import _split_authority
assert _split_authority("https://10.0.0.5:5120") == ("10.0.0.5", 5120)
assert _split_authority("localhost:5120") == ("localhost", 5120)
assert _split_authority(":5120") == ("localhost", 5120)
def test_split_authority_strips_ipv6_brackets() -> None:
from zb_mom_ww_mxgateway.options import _split_authority
# Bracketed IPv6 with port — brackets must be removed for ssl.get_server_certificate
assert _split_authority("[::1]:5120") == ("::1", 5120)
# Bare bracketed IPv6 (no port) — default port 443
assert _split_authority("[::1]") == ("::1", 443)
# Scheme-prefixed bracketed IPv6
assert _split_authority("grpc://[::1]:5120") == ("::1", 5120)
def test_tofu_connect_failure_raises_transport_error() -> None:
"""A failed cert pre-fetch surfaces the client's transport error type."""
options = ClientOptions(endpoint=f"127.0.0.1:{_free_port()}")
with pytest.raises(MxGatewayTransportError) as excinfo:
create_channel(options)
assert options.endpoint in str(excinfo.value)
def test_require_certificate_validation_uses_system_trust() -> None:
"""``require_certificate_validation`` must not attempt a TOFU pre-fetch."""
# Pointing at a closed port: with system-trust the channel is created lazily
# (no eager pre-fetch), so create_channel must succeed without connecting.
options = ClientOptions(
endpoint=f"127.0.0.1:{_free_port()}",
require_certificate_validation=True,
)
channel = create_channel(options)
assert isinstance(channel, grpc.aio.Channel)
-13
View File
@@ -76,19 +76,6 @@ types.
cargo run -p mxgw-cli -- smoke --endpoint https://mxgateway.example.local:5001 --tls --ca-file C:\certs\mxgateway-ca.pem --server-name-override mxgateway.example.local --api-key-env MXGATEWAY_API_KEY --item TestChildObject.TestInt --json
```
### TLS trust (pin-only)
The gateway can auto-generate its own self-signed certificate (it has no PKI).
Unlike the other clients, the Rust client is **not** lenient: tonic 0.13.1
exposes no public hook to inject a custom certificate verifier, so TLS over Rust
is pin-only. A TLS connection requires either `--ca-file` /
`ClientOptions::with_ca_file(...)` to pin a CA (export the gateway's self-signed
certificate and pin it), or `--require-certificate-validation` /
`with_require_certificate_validation(true)` to verify against the system trust
roots. TLS with neither set fails `connect` with a clear, actionable error rather
than accepting the certificate. See
[Gateway Configuration](../../docs/GatewayConfiguration.md#automatic-self-signed-certificate).
## Library Surface
`ClientOptions` configures endpoint, API key, plaintext or TLS transport,
-19
View File
@@ -189,25 +189,6 @@ Support:
- custom CA file,
- domain override.
### Trust posture (pin-only)
The gateway can serve a self-signed certificate it generates itself (it has no
PKI). Rust is the **exception** to the lenient-by-default posture the other
clients use: tonic 0.13.1 exposes no public hook to inject a custom certificate
verifier, so the Rust client cannot accept an arbitrary certificate. TLS over the
Rust client is therefore **pin-only** — it requires either:
- `ClientOptions::with_ca_file(...)` to pin a CA (the supported path for the
gateway's self-signed certificate; export the certificate and pin it), or
- `ClientOptions::with_require_certificate_validation(true)` to verify against the
system trust roots.
With TLS enabled (`with_plaintext(false)`), no pinned CA, and certificate
validation not required, `GatewayClient::connect` rejects the connection with a
clear, actionable error pointing at `with_ca_file` /
`require_certificate_validation` rather than silently accepting the certificate.
The CLI exposes `--ca-file` and `--require-certificate-validation`.
## Streaming
Expose event streams as a `Stream<Item = Result<MxEvent, Error>>`. Dropping the
-8
View File
@@ -426,11 +426,6 @@ struct ConnectionArgs {
ca_file: Option<PathBuf>,
#[arg(long)]
server_name_override: Option<String>,
/// Verify the server certificate against the system trust roots even
/// without a pinned CA. The Rust client's default is to require a CA
/// file (see `--ca-file`); set this flag to use system roots instead.
#[arg(long)]
require_certificate_validation: bool,
#[arg(long, default_value_t = 10)]
connect_timeout_seconds: u64,
#[arg(long, default_value_t = 30)]
@@ -458,9 +453,6 @@ impl ConnectionArgs {
if let Some(server_name_override) = &self.server_name_override {
options = options.with_server_name_override(server_name_override);
}
if self.require_certificate_validation {
options = options.with_require_certificate_validation(true);
}
options
}
+16 -3
View File
@@ -6,8 +6,10 @@
//! code should prefer [`GatewayClient::open_session`] and the [`Session`]
//! handle it returns, rather than the `*_raw` methods.
use std::fs;
use tonic::codegen::InterceptedService;
use tonic::transport::Channel;
use tonic::transport::{Certificate, Channel, ClientTlsConfig};
use tonic::Request;
use crate::auth::AuthInterceptor;
@@ -19,7 +21,7 @@ use crate::generated::mxaccess_gateway::v1::{
OpenSessionReply, OpenSessionRequest, QueryActiveAlarmsRequest, StreamAlarmsRequest,
StreamEventsRequest,
};
use crate::options::{build_tls_config, ClientOptions};
use crate::options::ClientOptions;
use crate::session::Session;
/// Generated gateway client wrapped in the auth interceptor that
@@ -76,7 +78,18 @@ impl GatewayClient {
})?;
endpoint = endpoint.connect_timeout(options.connect_timeout());
if let Some(tls) = build_tls_config(&options)? {
if !options.plaintext() {
let mut tls = ClientTlsConfig::new();
if let Some(server_name) = options.server_name_override() {
tls = tls.domain_name(server_name.to_owned());
}
if let Some(ca_file) = options.ca_file() {
let certificate = fs::read(ca_file).map_err(|source| Error::InvalidEndpoint {
endpoint: options.endpoint().to_owned(),
detail: format!("failed to read CA file {}: {source}", ca_file.display()),
})?;
tls = tls.ca_certificate(Certificate::from_pem(certificate));
}
endpoint = endpoint.tls_config(tls)?;
}
+15 -3
View File
@@ -6,12 +6,13 @@
//! re-exported through [`crate::generated::galaxy_repository::v1`].
use std::collections::HashSet;
use std::fs;
use std::sync::Arc;
use prost_types::Timestamp;
use tokio::sync::Mutex as AsyncMutex;
use tonic::codegen::InterceptedService;
use tonic::transport::Channel;
use tonic::transport::{Certificate, Channel, ClientTlsConfig};
use tonic::Request;
use crate::auth::AuthInterceptor;
@@ -22,7 +23,7 @@ use crate::generated::galaxy_repository::v1::{
DiscoverHierarchyRequest, GalaxyObject, GetLastDeployTimeRequest, TestConnectionRequest,
WatchDeployEventsRequest,
};
use crate::options::{build_tls_config, ClientOptions};
use crate::options::ClientOptions;
const DISCOVER_HIERARCHY_PAGE_SIZE: i32 = 5000;
const BROWSE_CHILDREN_PAGE_SIZE: i32 = 500;
@@ -182,7 +183,18 @@ impl GalaxyClient {
})?;
endpoint = endpoint.connect_timeout(options.connect_timeout());
if let Some(tls) = build_tls_config(&options)? {
if !options.plaintext() {
let mut tls = ClientTlsConfig::new();
if let Some(server_name) = options.server_name_override() {
tls = tls.domain_name(server_name.to_owned());
}
if let Some(ca_file) = options.ca_file() {
let certificate = fs::read(ca_file).map_err(|source| Error::InvalidEndpoint {
endpoint: options.endpoint().to_owned(),
detail: format!("failed to read CA file {}: {source}", ca_file.display()),
})?;
tls = tls.ca_certificate(Certificate::from_pem(certificate));
}
endpoint = endpoint.tls_config(tls)?;
}
-94
View File
@@ -3,14 +3,10 @@
//! chain of `with_*` setters; the `Debug` impl redacts the API key.
use std::fmt;
use std::fs;
use std::path::PathBuf;
use std::time::Duration;
use tonic::transport::{Certificate, ClientTlsConfig};
use crate::auth::ApiKey;
use crate::error::Error;
const DEFAULT_MAX_GRPC_MESSAGE_BYTES: usize = 16 * 1024 * 1024;
@@ -26,7 +22,6 @@ pub struct ClientOptions {
api_key: Option<ApiKey>,
plaintext: bool,
ca_file: Option<PathBuf>,
require_certificate_validation: bool,
server_name_override: Option<String>,
connect_timeout: Duration,
call_timeout: Duration,
@@ -43,7 +38,6 @@ impl ClientOptions {
api_key: None,
plaintext: true,
ca_file: None,
require_certificate_validation: false,
server_name_override: None,
connect_timeout: Duration::from_secs(10),
call_timeout: Duration::from_secs(30),
@@ -73,22 +67,6 @@ impl ClientOptions {
self
}
/// Require TLS certificate verification even without a pinned CA. Default
/// false: the gateway's self-signed certificate is accepted (internal-tool
/// posture). Setting a CA file always verifies.
///
/// Note for Rust: tonic 0.13's `ClientTlsConfig` exposes no hook for a
/// custom rustls verifier, so the Rust client cannot accept an arbitrary
/// self-signed certificate the way the other clients do. With the default
/// (false) and no pinned CA, [`crate::client::GatewayClient::connect`]
/// rejects the TLS connection and asks for a CA file. Either pin a CA via
/// [`ClientOptions::with_ca_file`] (the supported lenient path on Rust) or
/// set this `true` to verify against the system trust roots.
pub fn with_require_certificate_validation(mut self, require: bool) -> Self {
self.require_certificate_validation = require;
self
}
/// Override the SNI/server name used during the TLS handshake. Useful
/// when the dial-target host name does not match the certificate.
pub fn with_server_name_override(mut self, server_name_override: impl Into<String>) -> Self {
@@ -143,12 +121,6 @@ impl ClientOptions {
self.ca_file.as_ref()
}
/// Whether TLS certificate verification is required even without a pinned
/// CA. See [`ClientOptions::with_require_certificate_validation`].
pub fn require_certificate_validation(&self) -> bool {
self.require_certificate_validation
}
/// Optional SNI / server-name override for TLS handshakes.
pub fn server_name_override(&self) -> Option<&str> {
self.server_name_override.as_deref()
@@ -175,68 +147,6 @@ impl ClientOptions {
}
}
/// Build the [`ClientTlsConfig`] for a non-plaintext connection described by
/// `options`, applying the lenient-default guard that is the **Rust
/// pin-only exception**.
///
/// Returns `Ok(None)` when `options.plaintext()` is `true` (no TLS needed).
/// Returns `Ok(Some(tls))` when a valid TLS config can be assembled.
/// Returns `Err(Error::InvalidEndpoint)` when TLS is requested but no pinned
/// CA was provided and `require_certificate_validation` is `false`.
///
/// # Why this guard exists
///
/// `tonic` 0.13's `ClientTlsConfig` builds its rustls verifier inside a
/// crate-private connector and exposes no hook for a custom
/// `ServerCertVerifier`. The Rust client therefore cannot accept an arbitrary
/// self-signed certificate the way the other language clients do. Rather than
/// silently falling back to system-root verification (which always fails
/// against a self-signed gateway certificate), we reject the configuration
/// early with an actionable error.
pub(crate) fn build_tls_config(options: &ClientOptions) -> Result<Option<ClientTlsConfig>, Error> {
if options.plaintext() {
return Ok(None);
}
let mut tls = ClientTlsConfig::new();
if let Some(server_name) = options.server_name_override() {
tls = tls.domain_name(server_name.to_owned());
}
if let Some(ca_file) = options.ca_file() {
let certificate = fs::read(ca_file).map_err(|source| Error::InvalidEndpoint {
endpoint: options.endpoint().to_owned(),
detail: format!("failed to read CA file {}: {source}", ca_file.display()),
})?;
tls = tls.ca_certificate(Certificate::from_pem(certificate));
} else if !options.require_certificate_validation() {
// Lenient-default fallback (Rust pin-only exception): tonic
// 0.13's `ClientTlsConfig` builds its rustls verifier inside a
// crate-private connector and exposes no hook for a custom
// `ServerCertVerifier`, so — unlike the other clients — the
// Rust client cannot accept an arbitrary self-signed cert. Pin
// the gateway's CA instead, or opt into strict verification
// against the system trust roots. We reject here rather than
// silently verifying against system roots (which would fail a
// self-signed gateway with a confusing handshake error).
//
// Note: a server-name override affects SNI (the hostname sent
// in the TLS ClientHello) but does NOT pin trust. Overriding
// the server name alone does not bypass certificate validation.
return Err(Error::InvalidEndpoint {
endpoint: options.endpoint().to_owned(),
detail: "TLS requested without a pinned CA. The Rust client cannot accept an \
arbitrary self-signed certificate (tonic 0.13 exposes no custom \
rustls verifier). Pin the gateway certificate with \
ClientOptions::with_ca_file, or call \
ClientOptions::with_require_certificate_validation(true) to verify \
against the system trust roots. Note: a server-name override \
affects SNI but does not pin trust."
.to_owned(),
});
}
Ok(Some(tls))
}
impl Default for ClientOptions {
fn default() -> Self {
Self::new("http://127.0.0.1:5000")
@@ -251,10 +161,6 @@ impl fmt::Debug for ClientOptions {
.field("api_key", &self.api_key.as_ref().map(|_| "<redacted>"))
.field("plaintext", &self.plaintext)
.field("ca_file", &self.ca_file)
.field(
"require_certificate_validation",
&self.require_certificate_validation,
)
.field("server_name_override", &self.server_name_override)
.field("connect_timeout", &self.connect_timeout)
.field("call_timeout", &self.call_timeout)
-137
View File
@@ -1,137 +0,0 @@
//! TLS posture coverage for the Rust client.
//!
//! tonic 0.13.1's `ClientTlsConfig` exposes no hook for a custom rustls
//! `ServerCertVerifier` (the verifier is built internally inside the
//! crate-private `TlsConnector`), so the Rust client cannot implement the
//! "accept any server certificate" lenient default the other clients use.
//! Rust is therefore the documented **pin-only exception**: TLS without a
//! pinned CA is rejected up front with a clear, actionable error, and
//! supplying a CA file is the supported path. These tests pin that contract.
use std::time::Duration;
use zb_mom_ww_mxgateway_client::{ClientOptions, Error, GalaxyClient, GatewayClient};
/// Drive `connect` to its error without requiring `GatewayClient: Debug`
/// (the success arm is dropped explicitly so `unwrap_err` is unnecessary).
async fn connect_err(options: ClientOptions) -> Error {
match GatewayClient::connect(options).await {
Ok(_client) => panic!("connect unexpectedly succeeded against a dead TLS address"),
Err(error) => error,
}
}
#[tokio::test]
async fn tls_without_ca_is_rejected_with_actionable_error_by_default() {
let options = ClientOptions::new("https://127.0.0.1:1")
.with_plaintext(false)
.with_connect_timeout(Duration::from_millis(200));
let error = connect_err(options).await;
let Error::InvalidEndpoint { detail, .. } = error else {
panic!("expected InvalidEndpoint, got {error:?}");
};
// The message must point the caller at the supported remedy (pin a CA)
// and name the opt-in escape hatch.
assert!(
detail.contains("ca_file") || detail.contains("CA"),
"error should instruct the user to pass a CA file: {detail}"
);
assert!(
detail.contains("require_certificate_validation"),
"error should mention the require_certificate_validation opt-in: {detail}"
);
}
#[tokio::test]
async fn tls_with_require_certificate_validation_does_not_short_circuit() {
// With strict verification opted in, the no-CA guard must not fire; the
// connect attempt instead proceeds to the transport (and fails to reach
// the dead address) rather than returning the "CA required" guard error.
let options = ClientOptions::new("https://127.0.0.1:1")
.with_plaintext(false)
.with_require_certificate_validation(true)
.with_connect_timeout(Duration::from_millis(200));
let error = connect_err(options).await;
assert!(
!matches!(&error, Error::InvalidEndpoint { detail, .. }
if detail.contains("require_certificate_validation")),
"strict verification must bypass the no-CA guard, got {error:?}"
);
}
#[tokio::test]
async fn tls_with_ca_file_is_permitted_and_proceeds_past_the_guard() {
// Pinning a CA is the supported TLS path: the no-CA guard must not fire.
// We hand it a readable PEM file; construction proceeds past the guard
// and only fails later at the transport (dead address / handshake).
let ca_path = std::env::temp_dir().join("mxgw-rust-tls-ca-fixture.pem");
std::fs::write(&ca_path, SELF_SIGNED_CA_PEM).unwrap();
let options = ClientOptions::new("https://127.0.0.1:1")
.with_plaintext(false)
.with_ca_file(&ca_path)
.with_connect_timeout(Duration::from_millis(200));
let error = connect_err(options).await;
let _ = std::fs::remove_file(&ca_path);
assert!(
!matches!(&error, Error::InvalidEndpoint { detail, .. }
if detail.contains("require_certificate_validation")),
"pinning a CA must bypass the no-CA guard, got {error:?}"
);
}
/// Drive `GalaxyClient::connect` to its error (mirrors `connect_err` above).
async fn galaxy_connect_err(options: ClientOptions) -> Error {
match GalaxyClient::connect(options).await {
Ok(_client) => {
panic!("GalaxyClient::connect unexpectedly succeeded against a dead TLS address")
}
Err(error) => error,
}
}
#[tokio::test]
async fn galaxy_tls_without_ca_is_rejected_with_actionable_error_by_default() {
// GalaxyClient::connect must apply the same TLS guard as GatewayClient —
// TLS without a pinned CA (and without require_certificate_validation)
// returns a clear, actionable InvalidEndpoint error.
let options = ClientOptions::new("https://127.0.0.1:1")
.with_plaintext(false)
.with_connect_timeout(Duration::from_millis(200));
let error = galaxy_connect_err(options).await;
let Error::InvalidEndpoint { detail, .. } = error else {
panic!("expected InvalidEndpoint, got {error:?}");
};
assert!(
detail.contains("ca_file") || detail.contains("CA"),
"error should instruct the user to pass a CA file: {detail}"
);
assert!(
detail.contains("require_certificate_validation"),
"error should mention the require_certificate_validation opt-in: {detail}"
);
}
/// A throwaway self-signed CA certificate (PEM). Only needs to parse as a
/// PEM trust root so the CA-pinning path is exercised past the guard.
const SELF_SIGNED_CA_PEM: &str = "-----BEGIN CERTIFICATE-----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-----END CERTIFICATE-----
";
-13
View File
@@ -51,19 +51,6 @@ The shared inputs are:
The commands in the matrix use `MXGATEWAY_API_KEY` through each CLI's
`api-key-env` flag. They must not embed bearer tokens or raw API keys.
### TLS variant
The matrix runs over plaintext (`h2c`) by default. A TLS variant exists but stays
a manual/opt-in run, consistent with the gate above, because it needs the gateway
started with an HTTPS endpoint (an `https://` `MXGATEWAY_ENDPOINT`) and each CLI
switched to its TLS flag (`--tls` / `-tls` / `--plaintext=false` /
`plaintext=False`). The clients are lenient by default and accept the gateway's
auto-generated self-signed certificate without extra trust setup, except the Rust
CLI, which is pin-only and needs `--ca-file` or `--require-certificate-validation`
(and Python uses trust-on-first-use). See
[Gateway Configuration — Automatic self-signed certificate](./GatewayConfiguration.md#automatic-self-signed-certificate)
and each client README for the per-client TLS flags.
## JSON Comparison
Every command in the matrix requests JSON output. A runner can compare the
-36
View File
@@ -375,42 +375,6 @@ deployment-heavy box, multiply per-session SQL connections, and complicate the
cold-start path. Wire-side laziness solves the actual pain (oversized gRPC
replies and a heavy DOM) without disturbing the materialization model.
## TLS Auto-Certificate and Lenient Client Trust
Decision: when a Kestrel `https://` endpoint is configured without a certificate
of its own (and no `Kestrel:Certificates:Default` is set), the gateway generates
and persists a self-signed certificate rather than failing to start. Clients
connecting over TLS without a pinned CA accept whatever certificate the server
presents by default; pinning a CA restores full verification.
Rationale: `mxaccessgw` is an internal tool with no PKI to issue or distribute
certificates. The prior behavior — an `https` endpoint with no certificate
fails at startup with Kestrel's opaque "no server certificate was specified"
error — pushed operators toward plaintext (`h2c`), exposing the API key and
request payloads on the wire. Auto-generating a long-lived, persisted, reused
certificate lets TLS "just work" with zero certificate management, while the
lenient client default means clients connect to that self-signed certificate
without a manual trust step. Both choices are deliberate, not oversights:
strict-by-default would force PKI work this tool does not warrant. Plaintext-only
deployments are untouched — no certificate or key material is written for them —
and an operator who supplies a real certificate transparently overrides the
generated one.
Two clients diverge from "accept any certificate" because their gRPC stacks lack
a per-channel skip-verify hook:
- Python uses trust-on-first-use: it fetches the server's presented certificate
over a separate unverified probe and pins it for the channel, and defaults the
SNI/target-name override to `localhost` (the generated certificate always
carries a `localhost` SAN).
- Rust is pin-only: tonic exposes no public hook to inject a custom certificate
verifier, so TLS over Rust requires either a pinned CA or an explicit opt-in to
system-trust verification; otherwise connecting returns a clear, actionable
error.
See [Gateway Configuration — Automatic self-signed certificate](./GatewayConfiguration.md#automatic-self-signed-certificate)
and the per-client READMEs for the as-built behavior.
## Later Revisit Items
These are explicit post-v1 revisit items, not open blockers:
-179
View File
@@ -229,185 +229,6 @@ behavior.
The alarm monitor is independent of client sessions: `AcknowledgeAlarm` and
`StreamAlarms` are session-less RPCs served by the monitor.
## Host Endpoints and Transport Security (Kestrel)
The listening endpoints are **not** part of the `MxGateway` section. The gateway
uses the stock ASP.NET Core host (`WebApplication.CreateBuilder`) with no
`ConfigureKestrel` call in code, so endpoints come entirely from the standard
`Kestrel` configuration section. On the deployed hosts these values are supplied
as NSSM environment variables (`Kestrel__Endpoints__...`), not from
`appsettings.json`.
Two named endpoints are bound:
| Endpoint name | Purpose | Protocol requirement |
|---|---|---|
| `Http` | Public gRPC API (sessions, invoke, events, Galaxy browse) | HTTP/2 |
| `Dashboard` | Blazor dashboard and SignalR hubs | HTTP/1.1 (HTTP/2 optional) |
Both endpoints share one routing pipeline; the names only select which TCP port
serves which traffic. The gRPC endpoint must negotiate **HTTP/2**, which drives
the protocol settings below.
### Plaintext (current deployments)
Both running hosts (`10.100.0.48` and `wonder-app-vd03`) serve the gRPC port in
**cleartext HTTP/2 (`h2c`)**. Because cleartext HTTP/2 has no ALPN to negotiate
the protocol, the gRPC endpoint must be pinned to `Http2` with prior knowledge:
```text
Kestrel__Endpoints__Http__Url=http://0.0.0.0:5120
Kestrel__Endpoints__Http__Protocols=Http2
Kestrel__Endpoints__Dashboard__Url=http://0.0.0.0:5130
```
In this mode all client↔gateway traffic — including the
`authorization: Bearer mxgw_...` API key and any `WriteSecured` / `AuthenticateUser`
payloads — crosses the network **unencrypted**. This is acceptable only on a
trusted/isolated network segment. Prefer TLS for anything else.
### TLS
To encrypt the gRPC channel, give the `Http` endpoint an `https://` URL and a
certificate. Over TLS, ALPN negotiates HTTP/2, so the explicit `Protocols=Http2`
pin is no longer required (the default `Http1AndHttp2` works for gRPC over TLS).
`appsettings.json` form:
```json
{
"Kestrel": {
"Endpoints": {
"Http": {
"Url": "https://0.0.0.0:5120",
"Certificate": {
"Path": "C:\\ProgramData\\MxGateway\\certs\\gateway.pfx",
"Password": "<pfx-password>"
}
},
"Dashboard": {
"Url": "https://0.0.0.0:5130",
"Certificate": {
"Path": "C:\\ProgramData\\MxGateway\\certs\\gateway.pfx",
"Password": "<pfx-password>"
}
}
}
}
}
```
Equivalent NSSM environment-variable form (how config is delivered on the hosts —
see [server deploy mechanics in the project notes]):
```text
Kestrel__Endpoints__Http__Url=https://0.0.0.0:5120
Kestrel__Endpoints__Http__Certificate__Path=C:\ProgramData\MxGateway\certs\gateway.pfx
Kestrel__Endpoints__Http__Certificate__Password=<pfx-password>
Kestrel__Endpoints__Dashboard__Url=https://0.0.0.0:5130
Kestrel__Endpoints__Dashboard__Certificate__Path=C:\ProgramData\MxGateway\certs\gateway.pfx
Kestrel__Endpoints__Dashboard__Certificate__Password=<pfx-password>
```
Certificate sourcing options (any standard ASP.NET Core form is accepted):
| Form | Keys |
|---|---|
| PFX file | `Certificate:Path` (+ `Certificate:Password` if encrypted) |
| PEM pair | `Certificate:Path` (cert) + `Certificate:KeyPath` (private key) |
| Windows cert store | `Certificate:Subject`, `Certificate:Store` (e.g. `My`), `Certificate:Location` (`LocalMachine`), `Certificate:AllowInvalid` |
The certificate's CN/SAN must cover the host name clients dial (or clients must
set a server-name override — see below). The dashboard endpoint can keep its own
certificate independent of the gRPC endpoint; pair this with
`MxGateway:Dashboard:RequireHttpsCookie` (`true`) for production HTTPS.
### Automatic self-signed certificate
`mxaccessgw` is an internal tool with no PKI to issue certificates, so requiring
an operator to supply one before TLS works pushed deployments toward plaintext.
To avoid that, the gateway fills in a self-signed certificate when an HTTPS
endpoint is configured without one.
**Trigger.** At startup the gateway inspects `Kestrel:Endpoints:*`. If any
endpoint has an `https://` URL and no `Certificate` subsection of its own, and no
`Kestrel:Certificates:Default` is set, the gateway generates (or loads) a
persisted self-signed certificate and wires it in as the HTTPS *default* via
`ConfigureHttpsDefaults`. All-plaintext deployments are untouched: when no HTTPS
endpoint is configured, no certificate or key material is generated or written.
**Generated certificate.** ECDSA P-256, `serverAuth` EKU, validity ≈
`ValidityYears` (default 10 years, with one day of clock-skew slack before
`notBefore`). SANs cover `localhost`, the machine name (and its FQDN when
resolvable), each entry in `AdditionalDnsNames`, and the loopback addresses
`127.0.0.1` and `::1`.
**`MxGateway:Tls:*` options.** All optional; the zero-config path needs none of
them.
| Option | Default | Purpose |
|---|---|---|
| `Tls:SelfSignedCertPath` | `C:\ProgramData\MxGateway\certs\gateway-selfsigned.pfx` | Where the generated certificate is persisted |
| `Tls:ValidityYears` | `10` | Lifetime of the generated certificate (validated 1100) |
| `Tls:AdditionalDnsNames` | `[]` | Extra DNS SANs (e.g. a load-balancer name) |
| `Tls:RegenerateIfExpired` | `true` | Replace an expired persisted certificate instead of failing |
`ValidityYears` is validated by `GatewayOptionsValidator` (range 1100); the
"HTTPS endpoint configured but no certificate available" fail-fast lives in the
bootstrap/provider, because the validator only sees the `MxGateway` section, not
`Kestrel:Endpoints`.
**Persistence.** The PFX is written with an **empty** export password — a random
in-memory password could not be reused across restarts, which the
persist-and-reuse model requires. The private key is instead protected at rest by
filesystem permissions: a restrictive ACL on Windows (SYSTEM + Administrators,
inherited ACEs stripped) on the `certs` directory and file, and mode `0600` on
non-Windows. The write is atomic (hardened temp file, then move). The persisted
certificate is reused across restarts (stable thumbprint, so CA-pinning clients
keep working) and regenerated only when it is missing, expired (and
`RegenerateIfExpired` is `true`), or unreadable/corrupt. If the directory is not
writable or the ACL cannot be applied, the gateway fails fast with a diagnostic
naming the path rather than falling back to an in-memory certificate.
**Logging.** On generate or load, the gateway logs the certificate thumbprint,
SAN list, and `notAfter` at Information. The PFX bytes, export password, and
private key are never logged.
**Operator override.** The generated certificate is only the HTTPS *default*. To
use a real certificate, configure one explicitly — either per endpoint via
`Kestrel:Endpoints:<name>:Certificate` (`Path`/`Subject`/`Thumbprint`, etc., as
in the table above) or globally via `Kestrel:Certificates:Default`. An
explicitly-configured certificate takes precedence, and the gateway then writes
no self-signed material.
### Client side
Each official client opts into TLS explicitly. For the .NET client
(`MxGatewayClientOptions`):
| Option | Effect |
|---|---|
| `UseTls` (default `false`) | Enables TLS. Requires an `https://` endpoint; an `https://` endpoint without `UseTls` fails validation, and vice versa. |
| `CaCertificatePath` | Pins a custom root (self-signed / private CA) using `CustomRootTrust` chain validation instead of the OS trust store; the .NET client also enforces the certificate hostname/SAN match on this path. |
| `RequireCertificateValidation` (default `false`) | Forces OS/system-trust verification on a TLS connection with no pinned CA. Leave `false` for the lenient default. |
| `ServerNameOverride` | SNI / certificate host name override when the dialed host differs from the certificate CN/SAN. |
To pair with the auto-generated self-signed certificate above, the clients are
**lenient by default**: a TLS connection with no pinned CA accepts whatever
certificate the gateway presents. Pin `CaCertificatePath` to verify, or set
`RequireCertificateValidation` to force system-trust verification without
pinning. The other language clients expose the equivalent options; the exact
behavior differs per stack — Python uses trust-on-first-use and Rust is pin-only.
See each client README for the as-built behavior.
### Gateway↔worker IPC
Transport security here applies only to the public gRPC channel. The
gateway↔worker link is a per-session **named pipe**
(`mxaccess-gateway-{gatewayPid}-{sessionId}`), not a network socket. It is not
TLS-encrypted and does not need to be: it never leaves the local Windows host and
is secured by the OS pipe ACL. See [Worker Frame Protocol](./WorkerFrameProtocol.md).
## Related Documentation
- [Gateway Process Detailed Design](./GatewayProcessDesign.md)
-18
View File
@@ -243,27 +243,9 @@ services.AddGrpc(options => options.Interceptors.Add<GatewayGrpcAuthorizationInt
Because the interceptor runs before any handler, `MxAccessGatewayService` can safely assume the call has been authorized and that `IGatewayRequestIdentityAccessor.Current` is populated. The handler's only responsibility is to read the identity for `OpenSession` so the session is owned by the authenticated principal; it does not perform any authorization checks of its own. See [Authorization](./Authorization.md) for the policy and identity model.
## Transport Security
The gRPC endpoint runs over HTTP/2, in cleartext (`h2c`) or TLS depending on the
Kestrel endpoint configuration. The current deployments serve it in cleartext, so
the API key and request payloads cross the network unencrypted. The endpoint,
protocol pinning, and TLS certificate configuration — plus the corresponding
client `UseTls` / `CaCertificatePath` options — are documented in
[Host Endpoints and Transport Security](./GatewayConfiguration.md#host-endpoints-and-transport-security-kestrel).
To make TLS usable without PKI, the gateway can auto-generate and persist a
self-signed certificate when an HTTPS endpoint is configured without one, and the
language clients are lenient by default — a TLS connection with no pinned CA
accepts the presented certificate (with per-stack nuances: Python is
trust-on-first-use, Rust is pin-only). See
[Automatic self-signed certificate](./GatewayConfiguration.md#automatic-self-signed-certificate)
and each client README for the as-built behavior.
## Related Documentation
- [Contracts](./Contracts.md)
- [Sessions](./Sessions.md)
- [Authorization](./Authorization.md)
- [Gateway Configuration](./GatewayConfiguration.md)
- [Gateway Process Design](./GatewayProcessDesign.md)
@@ -1,156 +0,0 @@
# Gateway TLS Auto-Certificate and Lenient Client Trust — Design
Date: 2026-06-01
Status: Approved (brainstorming), pending implementation plan
## Problem
The gateway can serve gRPC and the dashboard over TLS, but only if an operator
supplies a certificate via the Kestrel `https://` endpoint config. With no cert,
an `https` endpoint fails at startup with Kestrel's opaque "No server certificate
was specified" error. Both current deployments therefore run plaintext (`h2c`),
exposing the API key and request payloads on the wire.
`mxaccessgw` is an internal tool. The goal is for TLS to "just work" with zero PKI
management: the gateway fabricates its own long-lived certificate when an HTTPS
endpoint is configured without one, and clients accept whatever certificate is
presented unless an operator explicitly opts into pinning.
## Decisions
1. **Gateway = fill-missing-cert-only.** No new "enable TLS" switch. TLS is still
driven by configuring a Kestrel `https://` endpoint. New behavior: when an
HTTPS endpoint has no `Certificate` section, the gateway generates/loads a
persisted self-signed cert instead of failing. Plaintext-only hosts are
untouched — no certificate or key material is ever written for them.
2. **Persist & reuse.** The self-signed cert is saved as a PFX under
`C:\ProgramData\MxGateway\certs`, reused across restarts, regenerated only if
missing, expired, or unreadable. Stable thumbprint; survives restarts; any
CA-pinning client keeps working.
3. **Clients = lenient TLS, plaintext default.** When a client connects over TLS
without a pinned CA, it skips verification (accepts any cert). Pinning a CA file
restores full verification. The per-client connection default (mostly
plaintext/`http`) does not change — TLS is still opt-in via the endpoint scheme.
**Scope boundary:** the gateway↔worker named-pipe IPC is unchanged (local,
OS-secured by the pipe ACL). This work touches only the public gRPC/dashboard
transport and the five language clients.
## Gateway component
New type `SelfSignedCertificateProvider` in
`src/ZB.MOM.WW.MxGateway.Server/Security/Tls/`.
1. **Detect need.** Inspect `Kestrel:Endpoints:*` configuration at startup. If any
endpoint has an `https://` URL and no `Certificate` subsection, a default cert
is needed. If none do, the provider is a no-op (no file written).
2. **Load-or-create.** Look for the persisted PFX. If present, valid, and
unexpired, load it. Otherwise generate and persist.
3. **Generate.** `CertificateRequest` with **ECDSA P-256**, `notBefore = now - 1
day` (clock-skew slack), `notAfter = now + ValidityYears`. SANs: `DNS=localhost`,
`DNS=<MachineName>`, `DNS=<MachineName.FQDN>` when resolvable, plus
`IP=127.0.0.1` and `IP=::1`. Server-auth EKU.
4. **Persist securely.** Write the PFX with an **empty** export password (a random
in-memory password cannot be reused across restarts, which the persist-and-reuse
decision requires); protect the private key with a restrictive ACL (SYSTEM +
Administrators + service account) on the `certs` directory and file on Windows,
and `0600` on non-Windows; atomic write (temp + rename). After generating, the
cert is reloaded from the persisted PFX so Kestrel always serves the on-disk key.
5. **Wire into Kestrel.** In `GatewayApplication.CreateBuilder`, add
`builder.WebHost.ConfigureKestrel(o => o.ConfigureHttpsDefaults(h =>
h.ServerCertificate = cert))`. `ConfigureHttpsDefaults` supplies the cert only
for HTTPS endpoints that did not specify their own, so an operator-configured
`Kestrel:Endpoints:*:Certificate` transparently overrides it. One hook covers
both the gRPC and dashboard ports.
### New config block `MxGateway:Tls`
All optional; the zero-config path needs none of them.
| Option | Default | Purpose |
|---|---|---|
| `Tls:SelfSignedCertPath` | `C:\ProgramData\MxGateway\certs\gateway-selfsigned.pfx` | Where the generated cert lives |
| `Tls:ValidityYears` | `10` | Lifetime of the generated cert |
| `Tls:AdditionalDnsNames` | `[]` | Extra SANs (e.g. a load-balancer name) |
| `Tls:RegenerateIfExpired` | `true` | Auto-replace an expired persisted cert |
Validated by `GatewayOptionsValidator`: `ValidityYears` in 1100,
`SelfSignedCertPath` is a valid path shape when non-blank, and
`AdditionalDnsNames` entries are non-blank. (The "https endpoint exists but cert
path is blank" fail-fast lives in the bootstrap/provider, not the validator,
because the validator only sees the `MxGateway` section, not `Kestrel:Endpoints`.)
**Logging:** on generate/load, log thumbprint + SAN list + `notAfter` at
Information. Never log the PFX password or private key.
## Client lenient-TLS behavior
Uniform rule: **TLS on + no CA pinned ⇒ skip verification; CA pinned ⇒ full
verification.** No transport default changes. Each client also exposes an explicit
switch to force-disable leniency (strict-without-pinning) for the future.
| Client | Mechanism | Effort |
|---|---|---|
| .NET | In `CreateHttpHandler`, when `UseTls` and `CaCertificatePath` empty, set `SslOptions.RemoteCertificateValidationCallback = (_,_,_,_) => true`. CA path keeps existing custom-root validation. | trivial |
| Go | In `buildCredentials`, when TLS and no `CACertFile`/`TLSConfig`, use `tls.Config{InsecureSkipVerify: true, ServerName: override}`. | trivial |
| Java | grpc-netty-shaded 1.76.0 ships `InsecureTrustManagerFactory`. When TLS and no CA, build `GrpcSslContexts.forClient().trustManager(InsecureTrustManagerFactory.INSTANCE)`. | easy |
| Python | grpc-python has no per-channel skip-verify. Fetch the server leaf cert at connect via `ssl.get_server_certificate((host, port))`, pass it as `root_certificates` to `ssl_channel_credentials`, plus `grpc.ssl_target_name_override`. Effectively trusts what is presented (TOFU). | moderate, special-cased |
| Rust | tonic 0.13.1 + rustls (`tls-ring`). Implement a custom `rustls::client::danger::ServerCertVerifier` that accepts everything, build a `rustls::ClientConfig` via `.dangerous().with_custom_certificate_verifier(...)`, feed it to the channel. May require a custom hyper-rustls connector if `ClientTlsConfig` will not take a raw rustls config. **Needs an API spike.** | highest |
### Honesty caveats
- **Python** is not literally "ignore the cert"; it pins whatever the server
presents on first contact via a separate unverified TLS probe. For a self-signed
internal cert this is the intended outcome. Documented as a difference.
- **Rust** leniency depends on the tonic 0.13 TLS surface. If a custom verifier is
disproportionately invasive, the fallback is to require a CA file for Rust TLS
(pin-only) and document Rust as the exception.
## Error handling
Gateway:
- Cert dir not writable / ACL fails ⇒ fail fast at startup with a diagnostic naming
the path and required permission. No silent in-memory fallback.
- Persisted PFX corrupt/unreadable ⇒ warn, regenerate, overwrite.
- Persisted cert expired ⇒ regenerate if `RegenerateIfExpired` (default), else fail
fast instructing the operator to delete it or enable regeneration.
- HTTPS endpoint configured but generation disabled / path empty ⇒ validator
rejects at startup rather than letting Kestrel throw its opaque error.
Clients: surface unchanged. Skip-verify cannot itself raise. Python's pre-fetch
wraps connect failure into the existing connect-error type with the endpoint in the
message. Rust pin-only fallback surfaces the existing CA-file error.
## Documentation (same commit as source, per CLAUDE.md)
- `docs/GatewayConfiguration.md` — extend the TLS section: auto-generation, the
`MxGateway:Tls:*` block, persistence location/ACL, thumbprint logging, operator
override via `Kestrel:Endpoints:*:Certificate`.
- Each client README + `*ClientDesign.md` — "TLS is lenient by default; pin a CA to
verify," with Python TOFU and any Rust caveat noted.
- `docs/DesignDecisions.md` — record both posture choices and the why (internal
tool, no PKI) so they are not mistaken for an oversight.
## Testing
Gateway (`MxGateway.Tests`, no MXAccess):
- `SelfSignedCertificateProvider`: SANs, server-auth EKU, `notAfter ≈ now +
ValidityYears`, ECDSA P-256.
- Load-or-create: valid persisted PFX reused (same thumbprint); expired regenerates
when enabled; corrupt regenerates with a warning.
- Detection: HTTPS-without-cert engages; all-plaintext no-ops and writes no file;
endpoint with its own cert is not overridden.
- `GatewayOptionsValidator`: new `Tls:*` rules.
- Host integration: `Kestrel:Endpoints:Http:Url=https://127.0.0.1:0` builds and
binds (today it throws "no certificate specified").
Clients: each test project gets a lenient-TLS test against a throwaway self-signed
cert — connect with no CA succeeds; pinning a wrong CA fails (proves pinning still
verifies). Python exercises the pre-fetch path; mark opt-in if loopback timing is
flaky. Standard (non-live) tests; no MXAccess or external services.
Cross-language: add a TLS variant note to `docs/CrossLanguageSmokeMatrix.md`;
running the matrix over TLS stays manual/opt-in, consistent with the existing gate.
Per-component verification follows CLAUDE.md's source-update table (build + test
each touched component independently).
File diff suppressed because it is too large Load Diff
@@ -1,18 +0,0 @@
{
"planPath": "docs/plans/2026-06-01-gateway-cert-autogen-implementation.md",
"tasks": [
{"id": 1, "subject": "Task 1: Add TlsOptions config + bind into GatewayOptions", "status": "pending"},
{"id": 2, "subject": "Task 2: Validate MxGateway:Tls in GatewayOptionsValidator", "status": "pending", "blockedBy": [1]},
{"id": 3, "subject": "Task 3: SelfSignedCertificateProvider.GenerateCertificate", "status": "pending", "blockedBy": [1]},
{"id": 4, "subject": "Task 4: SelfSignedCertificateProvider.LoadOrCreate (persist/reuse/regenerate/ACL)", "status": "pending", "blockedBy": [3]},
{"id": 5, "subject": "Task 5: KestrelTlsInspector (detect HTTPS-without-cert)", "status": "pending"},
{"id": 6, "subject": "Task 6: Wire auto-cert into GatewayApplication.CreateBuilder", "status": "pending", "blockedBy": [1, 4, 5]},
{"id": 7, "subject": "Task 7: .NET client lenient TLS by default", "status": "pending"},
{"id": 8, "subject": "Task 8: Go client lenient TLS by default", "status": "pending"},
{"id": 9, "subject": "Task 9: Java client lenient TLS by default", "status": "pending"},
{"id": 10, "subject": "Task 10: Python client lenient TLS via TOFU pre-fetch", "status": "pending"},
{"id": 11, "subject": "Task 11: Rust client lenient TLS via rustls verifier (spike + fallback)", "status": "pending"},
{"id": 12, "subject": "Task 12: Documentation", "status": "pending", "blockedBy": [6, 7, 8, 9, 10, 11]}
],
"lastUpdated": "2026-06-01"
}
@@ -43,7 +43,4 @@ public sealed class GatewayOptions
/// behaviour (alarms disabled).
/// </summary>
public AlarmsOptions Alarms { get; init; } = new();
/// <summary>Gets self-signed TLS certificate auto-generation options.</summary>
public TlsOptions Tls { get; init; } = new();
}
@@ -26,7 +26,6 @@ public sealed class GatewayOptionsValidator : IValidateOptions<GatewayOptions>
ValidateDashboard(options.Dashboard, failures);
ValidateProtocol(options.Protocol, failures);
ValidateAlarms(options.Alarms, failures);
ValidateTls(options.Tls, failures);
return failures.Count == 0
? ValidateOptionsResult.Success
@@ -263,36 +262,6 @@ public sealed class GatewayOptionsValidator : IValidateOptions<GatewayOptions>
}
}
private const int MinimumCertValidityYears = 1;
private const int MaximumCertValidityYears = 100;
private static void ValidateTls(TlsOptions options, List<string> failures)
{
if (options.ValidityYears is < MinimumCertValidityYears or > MaximumCertValidityYears)
{
failures.Add(
$"MxGateway:Tls:ValidityYears must be between {MinimumCertValidityYears} and {MaximumCertValidityYears}.");
}
// The default is non-blank, so this only catches an explicitly-blanked path.
AddIfBlank(
options.SelfSignedCertPath,
"MxGateway:Tls:SelfSignedCertPath must not be blank.",
failures);
AddIfInvalidPath(
options.SelfSignedCertPath,
"MxGateway:Tls:SelfSignedCertPath must be a valid filesystem path.",
failures);
foreach (string dns in options.AdditionalDnsNames)
{
if (string.IsNullOrWhiteSpace(dns))
{
failures.Add("MxGateway:Tls:AdditionalDnsNames entries must be non-blank.");
}
}
}
private static void ValidateProtocol(ProtocolOptions options, List<string> failures)
{
if (options.WorkerProtocolVersion != GatewayContractInfo.WorkerProtocolVersion)
@@ -1,22 +0,0 @@
namespace ZB.MOM.WW.MxGateway.Server.Configuration;
/// <summary>
/// Options controlling the gateway's self-signed certificate auto-generation.
/// Only consulted when a Kestrel HTTPS endpoint is configured without its own
/// certificate; plaintext deployments never trigger generation.
/// </summary>
public sealed class TlsOptions
{
/// <summary>Path to the persisted self-signed PFX. Reused across restarts.</summary>
public string SelfSignedCertPath { get; init; } =
@"C:\ProgramData\MxGateway\certs\gateway-selfsigned.pfx";
/// <summary>Lifetime in years of a freshly generated certificate.</summary>
public int ValidityYears { get; init; } = 10;
/// <summary>Extra DNS SANs to embed (e.g. a load-balancer name).</summary>
public IReadOnlyList<string> AdditionalDnsNames { get; init; } = [];
/// <summary>Regenerate the persisted certificate when it has expired.</summary>
public bool RegenerateIfExpired { get; init; } = true;
}
@@ -0,0 +1,64 @@
using ZB.MOM.WW.Telemetry.Serilog;
namespace ZB.MOM.WW.MxGateway.Server.Diagnostics;
/// <summary>
/// Bridges the gateway's <see cref="GatewayLogRedactor"/> policy onto the shared
/// <see cref="ILogRedactor"/> seam consumed by <c>ZB.MOM.WW.Telemetry.Serilog</c>'s redaction
/// enricher. Applied to every Serilog log event before it reaches a sink, it masks the same
/// secrets the original MEL-scope path masked: API-key bearer tokens / client identities
/// (<c>mxgw_*</c>) and command values for credential-bearing MXAccess commands. All masking
/// decisions delegate to <see cref="GatewayLogRedactor"/> — this type adds no new policy.
/// </summary>
public sealed class GatewayLogRedactorAdapter : ILogRedactor
{
/// <summary>Property name carrying a client identity / authorization header value.</summary>
private const string ClientIdentityProperty = "ClientIdentity";
/// <summary>Property name carrying a raw authorization header value.</summary>
private const string AuthorizationProperty = "Authorization";
/// <summary>Property name carrying the MXAccess command method, used to gate value redaction.</summary>
private const string CommandMethodProperty = "CommandMethod";
/// <summary>Property name carrying a command payload value that may bear credentials.</summary>
private const string CommandValueProperty = "CommandValue";
/// <summary>
/// Masks any sensitive values in <paramref name="properties"/> in place using the shared
/// <see cref="GatewayLogRedactor"/> policy. Identity/authorization properties have their API-key
/// secret stripped; a command value is redacted when its associated command method bears
/// credentials.
/// </summary>
/// <param name="properties">The mutable log-event property dictionary for the current event.</param>
public void Redact(IDictionary<string, object?> properties)
{
ArgumentNullException.ThrowIfNull(properties);
RedactIdentity(properties, ClientIdentityProperty);
RedactIdentity(properties, AuthorizationProperty);
RedactCommandValue(properties);
}
private static void RedactIdentity(IDictionary<string, object?> properties, string propertyName)
{
if (properties.TryGetValue(propertyName, out object? value) && value is string identity)
{
properties[propertyName] = GatewayLogRedactor.RedactClientIdentity(identity);
}
}
private static void RedactCommandValue(IDictionary<string, object?> properties)
{
if (!properties.TryGetValue(CommandValueProperty, out object? value) || value is null)
{
return;
}
string? commandMethod = properties.TryGetValue(CommandMethodProperty, out object? method)
? method as string
: null;
properties[CommandValueProperty] = GatewayLogRedactor.RedactCommandValue(commandMethod, value);
}
}
@@ -1,20 +0,0 @@
using Microsoft.Extensions.Logging;
namespace ZB.MOM.WW.MxGateway.Server.Diagnostics;
public static class GatewayLoggerExtensions
{
/// <summary>Begins a gateway log scope with the specified scope properties.</summary>
/// <param name="logger">Logger used for diagnostic output.</param>
/// <param name="scope">Scope properties to apply.</param>
/// <returns>A disposable that ends the scope when disposed.</returns>
public static IDisposable? BeginGatewayScope(
this ILogger logger,
GatewayLogScope scope)
{
ArgumentNullException.ThrowIfNull(logger);
ArgumentNullException.ThrowIfNull(scope);
return logger.BeginScope(scope.ToDictionary());
}
}
@@ -1,4 +1,5 @@
using Microsoft.Extensions.Primitives;
using Serilog.Context;
namespace ZB.MOM.WW.MxGateway.Server.Diagnostics;
@@ -17,7 +18,12 @@ public static class GatewayRequestLoggingMiddlewareExtensions
/// <summary>Header name for the command method name.</summary>
public const string CommandMethodHeaderName = "x-command-method";
/// <summary>Adds gateway request logging scope middleware that reads correlation headers and redacts sensitive data.</summary>
/// <summary>
/// Adds gateway request logging middleware that reads the correlation headers and pushes them
/// as Serilog <see cref="LogContext"/> properties for the duration of the request. The pushed
/// properties (SessionId / WorkerProcessId / CorrelationId / CommandMethod / ClientIdentity)
/// are disposed when the request completes; the shared redaction enricher masks any secrets.
/// </summary>
/// <param name="app">Application builder.</param>
public static IApplicationBuilder UseGatewayRequestLoggingScope(this IApplicationBuilder app)
{
@@ -25,21 +31,56 @@ public static class GatewayRequestLoggingMiddlewareExtensions
return app.Use(async (context, next) =>
{
ILogger logger = context.RequestServices
.GetRequiredService<ILoggerFactory>()
.CreateLogger("MxGateway.Request");
using IDisposable? scope = logger.BeginGatewayScope(new GatewayLogScope(
GatewayLogScope scope = new(
SessionId: ReadHeader(context, SessionIdHeaderName),
WorkerProcessId: ReadInt32Header(context, WorkerProcessIdHeaderName),
CorrelationId: ReadUInt64Header(context, CorrelationIdHeaderName),
CommandMethod: ReadHeader(context, CommandMethodHeaderName),
ClientIdentity: ReadHeader(context, "authorization")));
ClientIdentity: ReadHeader(context, "authorization"));
using IDisposable correlationScope = PushCorrelationProperties(scope);
await next(context);
});
}
/// <summary>
/// Pushes the populated <paramref name="scope"/> properties onto the Serilog
/// <see cref="LogContext"/>, returning a single disposable that pops them all when the request
/// completes. Only the properties present in <see cref="GatewayLogScope.ToDictionary"/> (which
/// already applies the client-identity redaction policy) are pushed.
/// </summary>
/// <param name="scope">The correlation properties for the current request.</param>
/// <returns>A disposable that removes the pushed properties on disposal.</returns>
private static IDisposable PushCorrelationProperties(GatewayLogScope scope)
{
Stack<IDisposable> pushed = new();
foreach (KeyValuePair<string, object?> property in scope.ToDictionary())
{
pushed.Push(LogContext.PushProperty(property.Key, property.Value));
}
return new CorrelationPropertyScope(pushed);
}
/// <summary>
/// Disposes the pushed <see cref="LogContext"/> property bindings in reverse order, restoring
/// the ambient context to its pre-request state.
/// </summary>
private sealed class CorrelationPropertyScope(Stack<IDisposable> bindings) : IDisposable
{
private readonly Stack<IDisposable> _bindings = bindings;
public void Dispose()
{
while (_bindings.Count > 0)
{
_bindings.Pop().Dispose();
}
}
}
private static string? ReadHeader(HttpContext context, string headerName)
{
return context.Request.Headers.TryGetValue(headerName, out StringValues values)
@@ -1,7 +1,5 @@
using System.Security.Cryptography.X509Certificates;
using Microsoft.AspNetCore.Hosting.StaticWebAssets;
using Microsoft.Extensions.Logging;
using Microsoft.Extensions.Logging.Configuration;
using Serilog;
using ZB.MOM.WW.MxGateway.Contracts;
using ZB.MOM.WW.MxGateway.Server.Alarms;
using ZB.MOM.WW.MxGateway.Server.Configuration;
@@ -14,6 +12,7 @@ using ZB.MOM.WW.MxGateway.Server.Security.Authentication;
using ZB.MOM.WW.MxGateway.Server.Security.Authorization;
using ZB.MOM.WW.MxGateway.Server.Sessions;
using ZB.MOM.WW.MxGateway.Server.Workers;
using ZB.MOM.WW.Telemetry.Serilog;
namespace ZB.MOM.WW.MxGateway.Server;
@@ -34,7 +33,10 @@ public static class GatewayApplication
WebApplicationBuilder builder = CreateBuilder(args);
WebApplication app = builder.Build();
// Push the per-request correlation properties (via Serilog LogContext) before the
// request-logging middleware emits its completion event, so those properties appear on it.
app.UseGatewayRequestLoggingScope();
app.UseSerilogRequestLogging();
app.UseStaticFiles();
app.UseAuthentication();
app.UseAuthorization();
@@ -58,7 +60,7 @@ public static class GatewayApplication
});
StaticWebAssetsLoader.UseStaticWebAssets(builder.Environment, builder.Configuration);
ConfigureSelfSignedTls(builder);
ConfigureSerilog(builder);
builder.Services.AddGatewayConfiguration();
builder.Services.AddSqliteAuthStore();
@@ -77,32 +79,28 @@ public static class GatewayApplication
return builder;
}
private static void ConfigureSelfSignedTls(WebApplicationBuilder builder)
/// <summary>
/// Replaces the default Microsoft.Extensions.Logging provider with the shared
/// <c>ZB.MOM.WW.Telemetry.Serilog</c> bootstrap (<see cref="ZbSerilogExtensions.AddZbSerilog"/>).
/// Sinks and minimum level come from the <c>Serilog</c> configuration section; identity
/// (<c>SiteId</c>/<c>NodeRole</c>) is read from <c>MxGateway:Telemetry</c> when present.
/// Also registers the project's <see cref="ILogRedactor"/> adapter so the shared redaction
/// enricher masks gateway secrets on every event.
/// </summary>
/// <param name="builder">The web application builder being configured.</param>
private static void ConfigureSerilog(WebApplicationBuilder builder)
{
if (!Security.Tls.KestrelTlsInspector.RequiresGeneratedCertificate(builder.Configuration))
{
return;
}
string? siteId = builder.Configuration["MxGateway:Telemetry:SiteId"];
string? nodeRole = builder.Configuration["MxGateway:Telemetry:NodeRole"];
Configuration.TlsOptions tlsOptions =
builder.Configuration.GetSection("MxGateway:Tls").Get<Configuration.TlsOptions>()
?? new Configuration.TlsOptions();
builder.Services.AddSingleton<ILogRedactor, GatewayLogRedactorAdapter>();
using ILoggerFactory loggerFactory = LoggerFactory.Create(logging =>
builder.AddZbSerilog(options =>
{
logging.AddConfiguration(builder.Configuration.GetSection("Logging"));
logging.AddConsole();
options.ServiceName = "mxgateway";
options.SiteId = string.IsNullOrWhiteSpace(siteId) ? null : siteId;
options.NodeRole = string.IsNullOrWhiteSpace(nodeRole) ? null : nodeRole;
});
Security.Tls.SelfSignedCertificateProvider provider = new(
tlsOptions,
loggerFactory.CreateLogger<Security.Tls.SelfSignedCertificateProvider>(),
TimeProvider.System);
X509Certificate2 certificate = provider.LoadOrCreate();
builder.WebHost.ConfigureKestrel(options =>
// The certificate is intentionally owned by Kestrel for the application
// lifetime; it is not disposed here.
options.ConfigureHttpsDefaults(https => https.ServerCertificate = certificate));
}
private static string ResolveContentRootPath()
@@ -1,51 +0,0 @@
using Microsoft.Extensions.Configuration;
namespace ZB.MOM.WW.MxGateway.Server.Security.Tls;
/// <summary>
/// Inspects the Kestrel configuration to decide whether the gateway must supply
/// a generated default certificate (an HTTPS endpoint exists with no certificate
/// of its own).
/// </summary>
public static class KestrelTlsInspector
{
/// <summary>
/// Returns <see langword="true"/> when at least one HTTPS endpoint in
/// <c>Kestrel:Endpoints</c> has no certificate of its own (no
/// <c>Certificate:Path</c>, <c>Certificate:Subject</c>, or
/// <c>Certificate:Thumbprint</c>), meaning the gateway must supply a
/// generated fallback certificate.
/// </summary>
public static bool RequiresGeneratedCertificate(IConfiguration configuration)
{
// A Kestrel default certificate applies to every endpoint that lacks its own.
// If the operator configured one, the gateway must not override it.
if (HasConfiguredCertificate(configuration.GetSection("Kestrel:Certificates:Default")))
{
return false;
}
IConfigurationSection endpoints = configuration.GetSection("Kestrel:Endpoints");
foreach (IConfigurationSection endpoint in endpoints.GetChildren())
{
string? url = endpoint["Url"];
if (string.IsNullOrWhiteSpace(url) ||
!url.StartsWith("https://", StringComparison.OrdinalIgnoreCase))
{
continue;
}
if (!HasConfiguredCertificate(endpoint.GetSection("Certificate")))
{
return true;
}
}
return false;
}
private static bool HasConfiguredCertificate(IConfigurationSection certificate)
=> !string.IsNullOrWhiteSpace(certificate["Path"]) ||
!string.IsNullOrWhiteSpace(certificate["Subject"]) ||
!string.IsNullOrWhiteSpace(certificate["Thumbprint"]);
}
@@ -1,239 +0,0 @@
using System.Net;
using System.Net.Sockets;
using System.Security.Cryptography;
using System.Security.Cryptography.X509Certificates;
using Microsoft.Extensions.Logging;
using ZB.MOM.WW.MxGateway.Server.Configuration;
namespace ZB.MOM.WW.MxGateway.Server.Security.Tls;
/// <summary>
/// Generates and persists a long-lived self-signed certificate used as the
/// Kestrel HTTPS default when no operator certificate is configured.
/// </summary>
public sealed class SelfSignedCertificateProvider
{
private const string ServerAuthOid = "1.3.6.1.5.5.7.3.1";
private const string SubjectAltNameOid = "2.5.29.17";
private readonly TlsOptions _options;
private readonly ILogger<SelfSignedCertificateProvider> _logger;
private readonly TimeProvider _timeProvider;
public SelfSignedCertificateProvider(
TlsOptions options,
ILogger<SelfSignedCertificateProvider> logger,
TimeProvider timeProvider)
{
_options = options;
_logger = logger;
_timeProvider = timeProvider;
}
/// <summary>Creates a fresh in-memory ECDSA P-256 self-signed certificate.</summary>
public X509Certificate2 GenerateCertificate()
{
using ECDsa key = ECDsa.Create(ECCurve.NamedCurves.nistP256);
CertificateRequest request = new(
new X500DistinguishedName("CN=MxAccessGateway Self-Signed"),
key,
HashAlgorithmName.SHA256);
// End-entity (non-CA) certificate: RFC 5280 marks BasicConstraints critical only for CA certs.
request.CertificateExtensions.Add(new X509BasicConstraintsExtension(false, false, 0, false));
request.CertificateExtensions.Add(new X509KeyUsageExtension(
X509KeyUsageFlags.DigitalSignature,
critical: true));
request.CertificateExtensions.Add(new X509EnhancedKeyUsageExtension(
[new Oid(ServerAuthOid, "Server Authentication")],
critical: false));
SubjectAlternativeNameBuilder san = new();
san.AddDnsName("localhost");
string machine = Environment.MachineName;
if (!string.IsNullOrWhiteSpace(machine))
{
san.AddDnsName(machine);
}
// Best-effort: add the machine FQDN when it differs from the short name and "localhost".
// GetHostEntry may fail if DNS is unavailable; skip silently in that case.
try
{
string fqdn = Dns.GetHostEntry(machine).HostName;
if (!string.IsNullOrWhiteSpace(fqdn)
&& !fqdn.Equals("localhost", StringComparison.OrdinalIgnoreCase)
&& !fqdn.Equals(machine, StringComparison.OrdinalIgnoreCase))
{
san.AddDnsName(fqdn);
}
}
catch (SocketException) { /* DNS not resolvable — FQDN SAN is optional */ }
catch (ArgumentException) { /* invalid host name — skip */ }
foreach (string extra in _options.AdditionalDnsNames)
{
if (!string.IsNullOrWhiteSpace(extra))
{
san.AddDnsName(extra);
}
}
san.AddIpAddress(IPAddress.Loopback);
san.AddIpAddress(IPAddress.IPv6Loopback);
request.CertificateExtensions.Add(san.Build());
DateTimeOffset now = _timeProvider.GetUtcNow();
return request.CreateSelfSigned(now.AddDays(-1), now.AddYears(_options.ValidityYears));
}
/// <summary>Loads the persisted certificate, regenerating when missing,
/// expired (and allowed), or unreadable.</summary>
public X509Certificate2 LoadOrCreate()
{
string path = _options.SelfSignedCertPath;
if (string.IsNullOrWhiteSpace(path))
{
throw new InvalidOperationException(
"MxGateway:Tls:SelfSignedCertPath must be set when an HTTPS endpoint has no certificate.");
}
if (File.Exists(path))
{
try
{
X509Certificate2 existing = X509CertificateLoader.LoadPkcs12FromFile(path, password: null, KeyStorageFlags());
if (existing.NotAfter.ToUniversalTime() > _timeProvider.GetUtcNow().UtcDateTime)
{
Log("Loaded", existing);
return existing;
}
if (!_options.RegenerateIfExpired)
{
string notAfter = existing.NotAfter.ToUniversalTime().ToString("u");
existing.Dispose();
throw new InvalidOperationException(
$"Persisted gateway certificate at '{path}' expired on {notAfter} " +
"and MxGateway:Tls:RegenerateIfExpired is false.");
}
_logger.LogWarning(
"Persisted gateway certificate at {Path} expired on {NotAfter:u}; regenerating.",
path, existing.NotAfter.ToUniversalTime());
existing.Dispose();
}
catch (CryptographicException ex)
{
_logger.LogWarning(ex,
"Persisted gateway certificate at {Path} is unreadable; regenerating.", path);
}
catch (Exception ex) when (ex is UnauthorizedAccessException or IOException)
{
// A permission/IO error must fail fast: do not regenerate (which would
// overwrite the operator's file), surface the cause instead.
throw new InvalidOperationException(
$"Persisted gateway certificate at '{path}' could not be read; the gateway " +
"process lacks read access or the file is otherwise inaccessible.", ex);
}
}
return GenerateAndPersist(path);
}
private X509Certificate2 GenerateAndPersist(string path)
{
using X509Certificate2 generated = GenerateCertificate();
byte[] pfx = generated.Export(X509ContentType.Pkcs12);
try
{
string? directory = Path.GetDirectoryName(path);
if (!string.IsNullOrEmpty(directory))
{
Directory.CreateDirectory(directory);
}
// The private-key bytes must never touch a world-readable file. Create the
// temp file empty, harden its permissions, and only then write the PFX into
// the already-protected file. The temp path is in the same directory as the
// target so the Move is atomic and preserves the hardened DACL/mode.
string temp = path + ".tmp";
using (File.Create(temp)) { }
HardenPermissions(temp);
// Writing into an existing file truncates content but preserves its ACL/mode.
// If the write or move fails the hardened temp file (which may contain private-key
// material) must not be left on disk; delete it best-effort before rethrowing.
try
{
File.WriteAllBytes(temp, pfx);
File.Move(temp, path, overwrite: true);
HardenPermissions(path);
}
catch (Exception)
{
try { File.Delete(temp); } catch { /* best effort */ }
throw;
}
X509Certificate2 loaded = X509CertificateLoader.LoadPkcs12FromFile(path, password: null, KeyStorageFlags());
Log("Generated", loaded);
return loaded;
}
finally
{
// pfx holds raw private-key material; clear it as soon as it is no longer needed.
Array.Clear(pfx, 0, pfx.Length);
}
}
private static X509KeyStorageFlags KeyStorageFlags()
=> OperatingSystem.IsWindows()
? X509KeyStorageFlags.MachineKeySet | X509KeyStorageFlags.PersistKeySet | X509KeyStorageFlags.Exportable
: X509KeyStorageFlags.Exportable;
private static void HardenPermissions(string path)
{
if (OperatingSystem.IsWindows())
{
HardenWindowsAcl(path);
}
else
{
File.SetUnixFileMode(path, UnixFileMode.UserRead | UnixFileMode.UserWrite);
}
}
[System.Runtime.Versioning.SupportedOSPlatform("windows")]
private static void HardenWindowsAcl(string path)
{
FileInfo file = new(path);
System.Security.AccessControl.FileSecurity security = new();
security.SetAccessRuleProtection(isProtected: true, preserveInheritance: false);
foreach (System.Security.Principal.WellKnownSidType sid in new[]
{
System.Security.Principal.WellKnownSidType.LocalSystemSid,
System.Security.Principal.WellKnownSidType.BuiltinAdministratorsSid,
})
{
System.Security.Principal.SecurityIdentifier identifier = new(sid, null);
security.AddAccessRule(new System.Security.AccessControl.FileSystemAccessRule(
identifier,
System.Security.AccessControl.FileSystemRights.FullControl,
System.Security.AccessControl.AccessControlType.Allow));
}
file.SetAccessControl(security);
}
private void Log(string action, X509Certificate2 cert)
{
string sans = cert.Extensions
.FirstOrDefault(e => e.Oid?.Value == SubjectAltNameOid)?
.Format(false) ?? "(none)";
_logger.LogInformation(
"{Action} gateway self-signed certificate: thumbprint={Thumbprint}, notAfter={NotAfter:u}, sans={Sans}",
action, cert.Thumbprint, cert.NotAfter.ToUniversalTime(), sans);
}
}
@@ -15,6 +15,13 @@
<ItemGroup>
<ProjectReference Include="..\ZB.MOM.WW.MxGateway.Contracts\ZB.MOM.WW.MxGateway.Contracts.csproj" />
<!--
Shared structured-logging bootstrap (ZB.MOM.WW.Telemetry.Serilog) lives in the sibling
scadaproj workspace. Cross-repo ProjectReference: the referenced project resolves its own
Directory.Build.props / Directory.Packages.props from its own tree, so it does not perturb
this repo's build settings. It transitively brings the ZB.MOM.WW.Telemetry core package.
-->
<ProjectReference Include="..\..\..\scadaproj\ZB.MOM.WW.Telemetry\src\ZB.MOM.WW.Telemetry.Serilog\ZB.MOM.WW.Telemetry.Serilog.csproj" />
</ItemGroup>
</Project>
@@ -1,8 +1,10 @@
{
"Logging": {
"LogLevel": {
"Serilog": {
"MinimumLevel": {
"Default": "Information",
"Microsoft.AspNetCore": "Warning"
"Override": {
"Microsoft.AspNetCore": "Warning"
}
}
}
}
@@ -1,9 +1,30 @@
{
"Logging": {
"LogLevel": {
"Serilog": {
"Using": [
"Serilog.Sinks.Console",
"Serilog.Sinks.File"
],
"MinimumLevel": {
"Default": "Information",
"Microsoft.AspNetCore": "Warning"
}
"Override": {
"Microsoft.AspNetCore": "Warning"
}
},
"WriteTo": [
{
"Name": "Console",
"Args": {
"outputTemplate": "[{Timestamp:HH:mm:ss} {Level:u3}] [{NodeRole}/{NodeHostname}] {Message:lj} {Properties:j}{NewLine}{Exception}"
}
},
{
"Name": "File",
"Args": {
"path": "logs/mxgateway-.log",
"rollingInterval": "Day"
}
}
]
},
"AllowedHosts": "*",
"MxGateway": {
@@ -1,68 +0,0 @@
using Microsoft.Extensions.Options;
using ZB.MOM.WW.MxGateway.Server.Configuration;
namespace ZB.MOM.WW.MxGateway.Tests.Configuration;
public sealed class GatewayOptionsValidatorTests
{
// Constructs the minimal valid GatewayOptions by relying on each sub-option's
// design-default values; those defaults are validated separately in GatewayOptionsTests.
private static GatewayOptions ValidOptions() => new();
private static GatewayOptions CloneWithTls(GatewayOptions source, TlsOptions tls)
=> new()
{
Authentication = source.Authentication,
Ldap = source.Ldap,
Worker = source.Worker,
Sessions = source.Sessions,
Events = source.Events,
Dashboard = source.Dashboard,
Protocol = source.Protocol,
Alarms = source.Alarms,
Tls = tls,
};
[Fact]
public void Validate_Succeeds_WithDefaultTlsOptions()
{
ValidateOptionsResult result = new GatewayOptionsValidator().Validate(null, ValidOptions());
Assert.True(result.Succeeded);
}
[Fact]
public void Validate_Fails_WhenTlsValidityYearsOutOfRange()
{
GatewayOptions withBadTls = CloneWithTls(ValidOptions(), new TlsOptions { ValidityYears = 0 });
ValidateOptionsResult result = new GatewayOptionsValidator().Validate(null, withBadTls);
Assert.True(result.Failed);
Assert.Contains(result.Failures!, f => f.Contains("MxGateway:Tls:ValidityYears"));
}
[Fact]
public void Validate_Fails_WhenTlsValidityYearsTooLarge()
{
GatewayOptions withBadTls = CloneWithTls(ValidOptions(), new TlsOptions { ValidityYears = 101 });
ValidateOptionsResult result = new GatewayOptionsValidator().Validate(null, withBadTls);
Assert.True(result.Failed);
Assert.Contains(result.Failures!, f => f.Contains("MxGateway:Tls:ValidityYears"));
}
[Fact]
public void Validate_Fails_WhenAdditionalDnsNameBlank()
{
GatewayOptions options = CloneWithTls(ValidOptions(), new TlsOptions { AdditionalDnsNames = [" "] });
ValidateOptionsResult result = new GatewayOptionsValidator().Validate(null, options);
Assert.True(result.Failed);
Assert.Contains(result.Failures!, f => f.Contains("MxGateway:Tls:AdditionalDnsNames"));
}
[Fact]
public void Validate_Fails_WhenSelfSignedCertPathBlank()
{
GatewayOptions options = CloneWithTls(ValidOptions(), new TlsOptions { SelfSignedCertPath = " " });
ValidateOptionsResult result = new GatewayOptionsValidator().Validate(null, options);
Assert.True(result.Failed);
Assert.Contains(result.Failures!, f => f.Contains("MxGateway:Tls:SelfSignedCertPath must not be blank."));
}
}
@@ -1,39 +0,0 @@
using Microsoft.Extensions.Configuration;
using ZB.MOM.WW.MxGateway.Server.Configuration;
using Xunit;
namespace ZB.MOM.WW.MxGateway.Tests.Configuration;
public sealed class TlsOptionsBindingTests
{
[Fact]
public void Defaults_AreApplied_WhenSectionAbsent()
{
TlsOptions options = new();
Assert.Equal(10, options.ValidityYears);
Assert.True(options.RegenerateIfExpired);
Assert.Empty(options.AdditionalDnsNames);
Assert.False(string.IsNullOrWhiteSpace(options.SelfSignedCertPath));
}
[Fact]
public void Binds_FromMxGatewayTlsSection()
{
IConfiguration config = new ConfigurationBuilder()
.AddInMemoryCollection(new Dictionary<string, string?>
{
["MxGateway:Tls:ValidityYears"] = "5",
["MxGateway:Tls:SelfSignedCertPath"] = @"C:\tmp\gw.pfx",
["MxGateway:Tls:RegenerateIfExpired"] = "false",
["MxGateway:Tls:AdditionalDnsNames:0"] = "gw.internal",
})
.Build();
GatewayOptions options = config.GetSection(GatewayOptions.SectionName).Get<GatewayOptions>()!;
Assert.Equal(5, options.Tls.ValidityYears);
Assert.Equal(@"C:\tmp\gw.pfx", options.Tls.SelfSignedCertPath);
Assert.False(options.Tls.RegenerateIfExpired);
Assert.Equal("gw.internal", Assert.Single(options.Tls.AdditionalDnsNames));
}
}
@@ -0,0 +1,92 @@
using ZB.MOM.WW.MxGateway.Server.Diagnostics;
using ZB.MOM.WW.Telemetry.Serilog;
namespace ZB.MOM.WW.MxGateway.Tests.Diagnostics;
/// <summary>
/// Pins that <see cref="GatewayLogRedactorAdapter"/> applies the gateway's redaction policy through
/// the shared <see cref="ILogRedactor"/> seam — the same secrets the former MEL-scope path masked
/// must still be masked once events flow through the Serilog redaction enricher.
/// </summary>
public sealed class GatewayLogRedactorAdapterTests
{
private readonly ILogRedactor _redactor = new GatewayLogRedactorAdapter();
/// <summary>Verifies the client identity property has its API-key secret stripped in place.</summary>
[Fact]
public void Redact_StripsApiKeySecretFromClientIdentity()
{
Dictionary<string, object?> properties = new()
{
["ClientIdentity"] = "Bearer mxgw_operator01_super-secret",
};
_redactor.Redact(properties);
Assert.Equal("Bearer mxgw_operator01_[redacted]", properties["ClientIdentity"]);
Assert.DoesNotContain("super-secret", (string?)properties["ClientIdentity"]);
}
/// <summary>Verifies a raw authorization header property is redacted too.</summary>
[Fact]
public void Redact_StripsApiKeySecretFromAuthorizationProperty()
{
Dictionary<string, object?> properties = new()
{
["Authorization"] = "Bearer mxgw_admin_top-secret",
};
_redactor.Redact(properties);
Assert.Equal("Bearer mxgw_admin_[redacted]", properties["Authorization"]);
}
/// <summary>Verifies a command value is redacted for a credential-bearing command method.</summary>
[Fact]
public void Redact_RedactsCommandValueForCredentialBearingCommand()
{
Dictionary<string, object?> properties = new()
{
["CommandMethod"] = "WriteSecured",
["CommandValue"] = "credential-bearing-value",
};
_redactor.Redact(properties);
Assert.Equal(GatewayLogRedactor.RedactedValue, properties["CommandValue"]);
}
/// <summary>Verifies a command value is redacted by default (value logging disabled) for any command.</summary>
[Fact]
public void Redact_RedactsCommandValueByDefault()
{
Dictionary<string, object?> properties = new()
{
["CommandMethod"] = "Write",
["CommandValue"] = "plaintext-tag-value",
};
_redactor.Redact(properties);
Assert.Equal(GatewayLogRedactor.RedactedValue, properties["CommandValue"]);
}
/// <summary>Verifies non-sensitive properties are left untouched.</summary>
[Fact]
public void Redact_LeavesNonSensitivePropertiesUnchanged()
{
Dictionary<string, object?> properties = new()
{
["SessionId"] = "session-1",
["CorrelationId"] = (ulong)99,
["ClientIdentity"] = "Bearer plain-token-no-marker",
};
_redactor.Redact(properties);
Assert.Equal("session-1", properties["SessionId"]);
Assert.Equal((ulong)99, properties["CorrelationId"]);
// No mxgw_ marker — identity passes through unchanged.
Assert.Equal("Bearer plain-token-no-marker", properties["ClientIdentity"]);
}
}
@@ -1,72 +0,0 @@
using System.Net.Security;
using System.Net.Sockets;
using System.Security.Cryptography.X509Certificates;
using Microsoft.AspNetCore.Builder;
using Microsoft.AspNetCore.Hosting.Server;
using Microsoft.AspNetCore.Hosting.Server.Features;
using Microsoft.Extensions.DependencyInjection;
using ZB.MOM.WW.MxGateway.Server;
namespace ZB.MOM.WW.MxGateway.Tests.Gateway;
public sealed class GatewayTlsBootstrapTests
{
/// <summary>
/// Verifies that when a Kestrel HTTPS endpoint is configured without its own certificate,
/// the gateway supplies the generated self-signed certificate as the Kestrel HTTPS default.
/// The host must start and bind, and the certificate served on the TLS handshake must be the
/// gateway's generated cert (subject <c>CN=MxAccessGateway Self-Signed</c>) — not an ambient
/// ASP.NET Core development certificate. On a host with no dev cert installed, starting a
/// cert-less HTTPS endpoint throws "No server certificate was specified"; on a host that has a
/// trusted dev cert, Kestrel would otherwise serve that dev cert (<c>CN=localhost</c>), so the
/// subject assertion is what makes this test fail without the wiring on either kind of host.
/// </summary>
[Fact]
public async Task Host_ServesGeneratedCertificate_WhenHttpsEndpointHasNoCertificate()
{
string certDir = Directory.CreateTempSubdirectory().FullName;
try
{
Environment.SetEnvironmentVariable("Kestrel__Endpoints__Test__Url", "https://127.0.0.1:0");
Environment.SetEnvironmentVariable(
"MxGateway__Tls__SelfSignedCertPath", Path.Combine(certDir, "gw.pfx"));
WebApplication app = GatewayApplication.Build([]);
await app.StartAsync();
try
{
string servedSubject = await ReadServedCertificateSubjectAsync(app);
Assert.Contains("MxAccessGateway Self-Signed", servedSubject, StringComparison.Ordinal);
}
finally
{
await app.StopAsync();
await app.DisposeAsync();
}
}
finally
{
Environment.SetEnvironmentVariable("Kestrel__Endpoints__Test__Url", null);
Environment.SetEnvironmentVariable("MxGateway__Tls__SelfSignedCertPath", null);
Directory.Delete(certDir, recursive: true);
}
}
private static async Task<string> ReadServedCertificateSubjectAsync(WebApplication app)
{
IServerAddressesFeature addresses =
app.Services.GetRequiredService<IServer>().Features.Get<IServerAddressesFeature>()
?? throw new InvalidOperationException("Server addresses feature was not available.");
Uri endpoint = new(addresses.Addresses.First());
using TcpClient client = new();
await client.ConnectAsync(endpoint.Host, endpoint.Port);
using SslStream ssl = new(
client.GetStream(),
leaveInnerStreamOpen: false,
userCertificateValidationCallback: (_, _, _, _) => true);
await ssl.AuthenticateAsClientAsync("127.0.0.1");
return ssl.RemoteCertificate?.Subject ?? "(none)";
}
}
@@ -1,68 +0,0 @@
using Microsoft.Extensions.Configuration;
using ZB.MOM.WW.MxGateway.Server.Security.Tls;
using Xunit;
namespace ZB.MOM.WW.MxGateway.Tests.Security.Tls;
public sealed class KestrelTlsInspectorTests
{
private static IConfiguration Config(params (string Key, string Value)[] entries)
=> new ConfigurationBuilder()
.AddInMemoryCollection(entries.ToDictionary(e => e.Key, e => (string?)e.Value))
.Build();
[Fact]
public void RequiresGeneratedCertificate_True_WhenHttpsEndpointHasNoCertificate()
=> Assert.True(KestrelTlsInspector.RequiresGeneratedCertificate(
Config(("Kestrel:Endpoints:Http:Url", "https://0.0.0.0:5120"))));
[Fact]
public void RequiresGeneratedCertificate_False_WhenAllEndpointsPlaintext()
=> Assert.False(KestrelTlsInspector.RequiresGeneratedCertificate(
Config(("Kestrel:Endpoints:Http:Url", "http://0.0.0.0:5120"))));
[Fact]
public void RequiresGeneratedCertificate_False_WhenHttpsEndpointHasOwnCertificate()
=> Assert.False(KestrelTlsInspector.RequiresGeneratedCertificate(
Config(
("Kestrel:Endpoints:Http:Url", "https://0.0.0.0:5120"),
("Kestrel:Endpoints:Http:Certificate:Path", @"C:\certs\real.pfx"))));
[Fact]
public void RequiresGeneratedCertificate_False_WhenNoEndpointsConfigured()
=> Assert.False(KestrelTlsInspector.RequiresGeneratedCertificate(Config()));
[Fact]
public void RequiresGeneratedCertificate_False_WhenHttpsEndpointHasThumbprintOnly()
=> Assert.False(KestrelTlsInspector.RequiresGeneratedCertificate(
Config(
("Kestrel:Endpoints:Https:Url", "https://0.0.0.0:5120"),
("Kestrel:Endpoints:Https:Certificate:Thumbprint", "AABBCCDDEEFF00112233445566778899AABBCCDD"))));
[Fact]
public void RequiresGeneratedCertificate_False_WhenHttpsEndpointHasSubjectOnly()
=> Assert.False(KestrelTlsInspector.RequiresGeneratedCertificate(
Config(
("Kestrel:Endpoints:Https:Url", "https://0.0.0.0:5120"),
("Kestrel:Endpoints:Https:Certificate:Subject", "CN=myserver"))));
[Fact]
public void RequiresGeneratedCertificate_True_WhenHttpsUrlIsUppercase()
=> Assert.True(KestrelTlsInspector.RequiresGeneratedCertificate(
Config(("Kestrel:Endpoints:Https:Url", "HTTPS://0.0.0.0:5120"))));
[Fact]
public void RequiresGeneratedCertificate_False_WhenKestrelDefaultCertificateConfigured()
=> Assert.False(KestrelTlsInspector.RequiresGeneratedCertificate(
Config(
("Kestrel:Endpoints:Https:Url", "https://0.0.0.0:5120"),
("Kestrel:Certificates:Default:Path", @"C:\certs\default.pfx"))));
[Fact]
public void RequiresGeneratedCertificate_True_WhenMixedEndpointsAndOneHttpsHasNoCert()
=> Assert.True(KestrelTlsInspector.RequiresGeneratedCertificate(
Config(
("Kestrel:Endpoints:Grpc:Url", "https://0.0.0.0:5120"),
("Kestrel:Endpoints:Grpc:Certificate:Thumbprint", "AABBCCDDEEFF00112233445566778899AABBCCDD"),
("Kestrel:Endpoints:Dashboard:Url", "https://0.0.0.0:5130"))));
}
@@ -1,156 +0,0 @@
using System.Security.Cryptography.X509Certificates;
using Microsoft.Extensions.Logging.Abstractions;
using Microsoft.Extensions.Time.Testing;
using ZB.MOM.WW.MxGateway.Server.Configuration;
using ZB.MOM.WW.MxGateway.Server.Security.Tls;
using Xunit;
namespace ZB.MOM.WW.MxGateway.Tests.Security.Tls;
public sealed class SelfSignedCertificateProviderTests
{
private static SelfSignedCertificateProvider CreateProvider(TlsOptions options, FakeTimeProvider time)
=> new(options, NullLogger<SelfSignedCertificateProvider>.Instance, time);
[Fact]
public void GenerateCertificate_HasExpectedSansEkuAndValidity()
{
FakeTimeProvider time = new(new DateTimeOffset(2026, 1, 1, 0, 0, 0, TimeSpan.Zero));
TlsOptions options = new() { ValidityYears = 7, AdditionalDnsNames = ["gw.internal"] };
using X509Certificate2 cert = CreateProvider(options, time).GenerateCertificate();
Assert.Equal(time.GetUtcNow().AddYears(7).UtcDateTime.Date, cert.NotAfter.ToUniversalTime().Date);
Assert.True(cert.NotBefore.ToUniversalTime() < time.GetUtcNow().UtcDateTime);
Assert.True(cert.HasPrivateKey);
string sans = ReadSubjectAltNames(cert);
Assert.Contains("localhost", sans);
Assert.Contains("gw.internal", sans);
Assert.Contains(Environment.MachineName, sans);
// Format() renders IP SANs as "IP Address:<addr>"; the IPv6 loopback may appear
// as "::1" or its expanded form depending on the platform crypto library.
Assert.Contains("127.0.0.1", sans);
Assert.True(sans.Contains("::1") || sans.Contains("0:0:0:0:0:0:0:1"),
$"Expected IPv6 loopback in SANs but got: {sans}");
X509EnhancedKeyUsageExtension eku = cert.Extensions.OfType<X509EnhancedKeyUsageExtension>().Single();
Assert.Contains(eku.EnhancedKeyUsages.Cast<System.Security.Cryptography.Oid>(),
o => o.Value == "1.3.6.1.5.5.7.3.1"); // serverAuth
}
[Fact]
public void LoadOrCreate_GeneratesPersistsAndReuses_SameThumbprint()
{
string dir = Directory.CreateTempSubdirectory().FullName;
try
{
string path = Path.Combine(dir, "gw.pfx");
FakeTimeProvider time = new(new DateTimeOffset(2026, 1, 1, 0, 0, 0, TimeSpan.Zero));
TlsOptions options = new() { SelfSignedCertPath = path };
using X509Certificate2 first = CreateProvider(options, time).LoadOrCreate();
Assert.True(File.Exists(path));
using X509Certificate2 second = CreateProvider(options, time).LoadOrCreate();
Assert.Equal(first.Thumbprint, second.Thumbprint); // reused, not regenerated
}
finally { Directory.Delete(dir, recursive: true); }
}
[Fact]
public void LoadOrCreate_Regenerates_WhenPersistedCertExpired()
{
string dir = Directory.CreateTempSubdirectory().FullName;
try
{
string path = Path.Combine(dir, "gw.pfx");
FakeTimeProvider time = new(new DateTimeOffset(2026, 1, 1, 0, 0, 0, TimeSpan.Zero));
TlsOptions options = new() { SelfSignedCertPath = path, ValidityYears = 1 };
using X509Certificate2 first = CreateProvider(options, time).LoadOrCreate();
time.Advance(TimeSpan.FromDays(800)); // past 1-year validity
using X509Certificate2 second = CreateProvider(options, time).LoadOrCreate();
Assert.NotEqual(first.Thumbprint, second.Thumbprint);
}
finally { Directory.Delete(dir, recursive: true); }
}
[Fact]
public void LoadOrCreate_Regenerates_WhenPersistedFileCorrupt()
{
string dir = Directory.CreateTempSubdirectory().FullName;
try
{
string path = Path.Combine(dir, "gw.pfx");
File.WriteAllText(path, "not a pfx");
TlsOptions options = new() { SelfSignedCertPath = path };
using X509Certificate2 cert = CreateProvider(options, new FakeTimeProvider()).LoadOrCreate();
Assert.True(cert.HasPrivateKey);
}
finally { Directory.Delete(dir, recursive: true); }
}
[Fact]
public void LoadOrCreate_Throws_WhenExpiredAndRegenerateDisabled()
{
string dir = Directory.CreateTempSubdirectory().FullName;
try
{
string path = Path.Combine(dir, "gw.pfx");
FakeTimeProvider time = new(new DateTimeOffset(2026, 1, 1, 0, 0, 0, TimeSpan.Zero));
TlsOptions options = new() { SelfSignedCertPath = path, ValidityYears = 1, RegenerateIfExpired = false };
using (CreateProvider(options, time).LoadOrCreate()) { }
time.Advance(TimeSpan.FromDays(800));
Assert.Throws<InvalidOperationException>(() => CreateProvider(options, time).LoadOrCreate());
}
finally { Directory.Delete(dir, recursive: true); }
}
[Fact]
public void LoadOrCreate_Throws_WhenSelfSignedCertPathBlank()
{
TlsOptions options = new() { SelfSignedCertPath = " " };
Assert.Throws<InvalidOperationException>(
() => CreateProvider(options, new FakeTimeProvider()).LoadOrCreate());
}
/// <summary>
/// Verifies that GenerateAndPersist cleans up the hardened .tmp file when persist fails.
/// The failure is induced by setting SelfSignedCertPath to a path whose parent directory
/// is an existing regular file, causing Directory.CreateDirectory (or the subsequent write)
/// to throw an IOException/UnauthorizedAccessException.
/// </summary>
[Fact]
public void LoadOrCreate_DeletesTempFile_WhenPersistFails()
{
string outerDir = Directory.CreateTempSubdirectory().FullName;
try
{
// Create a regular file at what would be the parent directory of the cert path.
// Any attempt to create that "directory" or write files into it must fail.
string fileActingAsDir = Path.Combine(outerDir, "notadir");
File.WriteAllText(fileActingAsDir, "block");
// Point the cert path inside the regular file — Directory.CreateDirectory will
// throw because the parent path component is a file, not a directory.
string certPath = Path.Combine(fileActingAsDir, "gw.pfx");
string expectedTemp = certPath + ".tmp";
TlsOptions options = new() { SelfSignedCertPath = certPath };
Assert.ThrowsAny<Exception>(() => CreateProvider(options, new FakeTimeProvider()).LoadOrCreate());
// The .tmp file must not be left behind.
Assert.False(File.Exists(expectedTemp), $"Leaked temp file: {expectedTemp}");
}
finally { Directory.Delete(outerDir, recursive: true); }
}
private const string SubjectAltNameOid = "2.5.29.17";
private static string ReadSubjectAltNames(X509Certificate2 cert)
=> cert.Extensions
.First(e => e.Oid?.Value == SubjectAltNameOid)
.Format(false);
}
@@ -7,7 +7,6 @@
<ItemGroup>
<PackageReference Include="coverlet.collector" Version="6.0.4" />
<PackageReference Include="Microsoft.Extensions.TimeProvider.Testing" Version="10.0.0" />
<PackageReference Include="Microsoft.NET.Test.Sdk" Version="17.14.1" />
<PackageReference Include="xunit" Version="2.9.3" />
<PackageReference Include="xunit.runner.visualstudio" Version="3.1.4" />