feat(wire): route_to_call trace flag + child_execution_id + trace_events; route_to_set_attributes trace (additive proto + mapper)

Flow Debugger (design 2026-09-05 §5.2/§5.4): site_command.proto gains trace=7 on
RouteToCallRequestDto, child_execution_id=6 + repeated sitestream.AuditEventDto trace_events=7
on RouteToCallResponseDto, trace=6 on RouteToSetAttributesRequestDto (imports
Protos/sitestream.proto the way central_control.proto does). Mapper round-trips them; an
empty trace_events list maps to null so an untraced/older site reads as 'not traced'.
Vendored C# regenerated via docker/regen-proto.sh (csproj untouched); golden samples extended.

Claude-Session: https://claude.ai/code/session_01GSLkkHKz9JrAtmJjmQK3cP
This commit is contained in:
Joseph Doherty
2026-09-05 23:52:50 -04:00
parent 2c86adb782
commit f15d0db0ad
5 changed files with 792 additions and 483 deletions
@@ -2420,7 +2420,8 @@ public static class SiteCommandDtoMapper
ScriptName = request.ScriptName,
Parameters = LooseValueCodec.ToProtoMapOrNull(request.Parameters),
Timestamp = Ts(request.Timestamp),
ParentExecutionId = GuidToWire(request.ParentExecutionId)
ParentExecutionId = GuidToWire(request.ParentExecutionId),
Trace = request.Trace
};
}
@@ -2437,7 +2438,8 @@ public static class SiteCommandDtoMapper
dto.ScriptName,
LooseValueCodec.FromProtoMapOrNull(dto.Parameters),
Dto(dto.Timestamp),
GuidFromWire(dto.ParentExecutionId));
GuidFromWire(dto.ParentExecutionId),
dto.Trace);
}
/// <summary>Projects a <see cref="RouteToGetAttributesRequest"/> onto the wire.</summary>
@@ -2485,7 +2487,8 @@ public static class SiteCommandDtoMapper
CorrelationId = request.CorrelationId,
InstanceUniqueName = request.InstanceUniqueName,
Timestamp = Ts(request.Timestamp),
ParentExecutionId = GuidToWire(request.ParentExecutionId)
ParentExecutionId = GuidToWire(request.ParentExecutionId),
Trace = request.Trace
};
foreach (var (key, value) in request.AttributeValues)
@@ -2508,7 +2511,8 @@ public static class SiteCommandDtoMapper
dto.InstanceUniqueName,
dto.AttributeValues.ToDictionary(kv => kv.Key, kv => kv.Value),
Dto(dto.Timestamp),
GuidFromWire(dto.ParentExecutionId));
GuidFromWire(dto.ParentExecutionId),
dto.Trace);
}
/// <summary>Projects a <see cref="RouteToWaitForAttributeRequest"/> onto the wire.</summary>
@@ -2564,14 +2568,24 @@ public static class SiteCommandDtoMapper
{
ArgumentNullException.ThrowIfNull(response);
return new RouteToCallResponseDto
var dto = new RouteToCallResponseDto
{
CorrelationId = response.CorrelationId,
Success = response.Success,
ReturnValue = LooseValueCodec.ToProtoOrNull(response.ReturnValue),
ErrorMessage = OrEmpty(response.ErrorMessage),
Timestamp = Ts(response.Timestamp)
Timestamp = Ts(response.Timestamp),
// Flow Debugger (design 2026-09-05 §5.4): additive fields 6/7.
ChildExecutionId = GuidToWire(response.ChildExecutionId)
};
if (response.TraceEvents is { Count: > 0 } traceEvents)
{
foreach (var evt in traceEvents)
{
dto.TraceEvents.Add(AuditEventDtoMapper.ToDto(evt));
}
}
return dto;
}
/// <summary>Rehydrates a <see cref="RouteToCallResponse"/> from the wire.</summary>
@@ -2586,7 +2600,11 @@ public static class SiteCommandDtoMapper
dto.Success,
LooseValueCodec.FromProto(dto.ReturnValue),
NullIfEmpty(dto.ErrorMessage),
Dto(dto.Timestamp));
Dto(dto.Timestamp),
GuidFromWire(dto.ChildExecutionId),
// An untraced run — or an older site that never sets the field — sends no
// events, which maps to null so the UI can label the lane "site not traced".
dto.TraceEvents.Count == 0 ? null : dto.TraceEvents.Select(AuditEventDtoMapper.FromDto).ToList());
}
/// <summary>Projects a <see cref="RouteToGetAttributesResponse"/> onto the wire.</summary>
@@ -5,6 +5,9 @@ package scadabridge.sitecommand.v1;
import "google/protobuf/timestamp.proto";
import "google/protobuf/duration.proto";
import "google/protobuf/wrappers.proto";
// Flow Debugger (design 2026-09-05 §5.4): RouteToCallResponseDto carries the site's
// ScriptTrace rows inline as the AuditEventDto sitestream.proto already defines.
import "Protos/sitestream.proto";
// ─────────────────────────────────────────────────────────────────────────────
// Site command plane (central → site) — the gRPC replacement for the per-site
@@ -765,6 +768,9 @@ message RouteToCallRequestDto {
LooseValueMap parameters = 4; // absent => null (distinct from empty)
google.protobuf.Timestamp timestamp = 5;
string parent_execution_id = 6; // empty string represents null
// Flow Debugger trace flag (design 2026-09-05 §5.2). proto3 default false = untraced,
// which is what an older central sends and what an older site reads.
bool trace = 7;
}
message RouteToCallResponseDto {
@@ -773,6 +779,12 @@ message RouteToCallResponseDto {
LooseValue return_value = 3; // absent => null
string error_message = 4; // empty string represents null
google.protobuf.Timestamp timestamp = 5;
// Flow Debugger (design §5.4): the routed site run's ExecutionId ("D" GUID form;
// empty represents null — an older site never sets it).
string child_execution_id = 6;
// Flow Debugger (design §5.4): the site's ScriptTrace audit events for a traced run,
// in emission order. Empty from an untraced run or an older site (mapper => null).
repeated sitestream.AuditEventDto trace_events = 7;
}
message RouteToGetAttributesRequestDto {
@@ -797,6 +809,8 @@ message RouteToSetAttributesRequestDto {
map<string, string> attribute_values = 3;
google.protobuf.Timestamp timestamp = 4;
string parent_execution_id = 5; // empty string represents null
// Flow Debugger trace flag (design 2026-09-05 §5.2/§5.3): arms the echo stamp.
bool trace = 6;
}
message RouteToSetAttributesResponseDto {
File diff suppressed because it is too large Load Diff
@@ -0,0 +1,95 @@
using ZB.MOM.WW.ScadaBridge.Commons.Messages.InboundApi;
using ZB.MOM.WW.ScadaBridge.Commons.Types.Audit;
using ZB.MOM.WW.ScadaBridge.Commons.Types.Debugging;
using ZB.MOM.WW.ScadaBridge.Commons.Types.Enums;
using ZB.MOM.WW.ScadaBridge.Communication.Grpc;
namespace ZB.MOM.WW.ScadaBridge.Communication.Tests;
/// <summary>
/// Flow Debugger (design 2026-09-05 §5.2/§5.4): the site_command.proto trace fields — the
/// <c>trace</c> flag on both route requests, and <c>child_execution_id</c> +
/// <c>trace_events</c> on the call response — round-trip through
/// <see cref="SiteCommandDtoMapper"/>, and an old-shape DTO (no fields set) maps to the
/// untraced defaults.
/// </summary>
public class SiteCommandDtoMapperTraceTests
{
private static readonly DateTimeOffset T = new(2026, 9, 5, 12, 0, 0, TimeSpan.Zero);
[Fact]
public void RouteToCallResponse_RoundTrips_ChildExecutionId_AndTraceEvents()
{
var child = Guid.NewGuid();
var run = ScadaBridgeAuditEventFactory.Create(AuditChannel.ScriptTrace, AuditKind.ScriptRun, AuditStatus.Delivered,
occurredAtUtc: T.UtcDateTime, actor: "ScriptActor:S", target: "S", sourceNode: "node-b",
executionId: child, parentExecutionId: Guid.NewGuid(), sourceSiteId: "site-a", durationMs: 40,
requestSummary: "{\"x\":1}", responseSummary: "\"ok\"",
extra: new ScriptTraceExtra { TriggerKind = "call" }.ToJson());
var write = ScadaBridgeAuditEventFactory.Create(AuditChannel.ScriptTrace, AuditKind.AttributeWrite, AuditStatus.Delivered,
occurredAtUtc: T.UtcDateTime.AddMilliseconds(5), actor: "ScriptActor:S", target: "A", executionId: child,
extra: new ScriptTraceExtra { Writes = [new AttributeWriteDetail("A", "1", true, null)] }.ToJson());
var original = new RouteToCallResponse("c", true, 7L, null, T, child, [run, write]);
var restored = SiteCommandDtoMapper.FromProto(SiteCommandDtoMapper.ToProto(original));
Assert.Equal(child, restored.ChildExecutionId);
Assert.NotNull(restored.TraceEvents);
Assert.Equal(2, restored.TraceEvents!.Count);
Assert.Equal(run.Action, restored.TraceEvents[0].Action);
Assert.Equal(run.DetailsJson, restored.TraceEvents[0].DetailsJson);
Assert.Equal(write.EventId, restored.TraceEvents[1].EventId);
Assert.Equal(original, restored with { TraceEvents = original.TraceEvents });
Assert.Equal(run, restored.TraceEvents[0]);
}
[Fact]
public void OldShapeDtos_MapToUntracedDefaults()
{
// A DTO from a site/central that predates the fields: proto3 defaults everywhere.
var call = SiteCommandDtoMapper.FromProto(new RouteToCallRequestDto
{
CorrelationId = "c", InstanceUniqueName = "i", ScriptName = "s",
Timestamp = Google.Protobuf.WellKnownTypes.Timestamp.FromDateTimeOffset(T),
});
Assert.False(call.Trace);
var set = SiteCommandDtoMapper.FromProto(new RouteToSetAttributesRequestDto
{
CorrelationId = "c", InstanceUniqueName = "i",
Timestamp = Google.Protobuf.WellKnownTypes.Timestamp.FromDateTimeOffset(T),
});
Assert.False(set.Trace);
var reply = SiteCommandDtoMapper.FromProto(new RouteToCallResponseDto
{
CorrelationId = "c", Success = true,
Timestamp = Google.Protobuf.WellKnownTypes.Timestamp.FromDateTimeOffset(T),
});
Assert.Null(reply.ChildExecutionId);
Assert.Null(reply.TraceEvents);
}
[Fact]
public void TraceFlag_PropagatesOnBothRequests()
{
var call = new RouteToCallRequest("c", "i", "s", null, T, Guid.NewGuid(), Trace: true);
Assert.True(SiteCommandDtoMapper.ToProto(call).Trace);
Assert.True(SiteCommandDtoMapper.FromProto(SiteCommandDtoMapper.ToProto(call)).Trace);
var set = new RouteToSetAttributesRequest("c", "i", new Dictionary<string, string> { ["a"] = "1" }, T, Guid.NewGuid(), Trace: true);
Assert.True(SiteCommandDtoMapper.ToProto(set).Trace);
Assert.True(SiteCommandDtoMapper.FromProto(SiteCommandDtoMapper.ToProto(set)).Trace);
}
[Fact]
public void EmptyTraceEventsList_MapsToNull_NotEmpty()
{
// An untraced run sends none; the CLR side distinguishes "none" (null) so the UI can
// label the site lane — an empty list would read as "traced, did nothing".
var original = new RouteToCallResponse("c", true, null, null, T, Guid.NewGuid(), Array.Empty<ZB.MOM.WW.Audit.AuditEvent>());
var restored = SiteCommandDtoMapper.FromProto(SiteCommandDtoMapper.ToProto(original));
Assert.Null(restored.TraceEvents);
Assert.Equal(original.ChildExecutionId, restored.ChildExecutionId);
}
}
@@ -1,3 +1,4 @@
using ZB.MOM.WW.Audit;
using ZB.MOM.WW.ScadaBridge.Commons.Interfaces.Protocol;
using ZB.MOM.WW.ScadaBridge.Commons.Messages.Artifacts;
using ZB.MOM.WW.ScadaBridge.Commons.Messages.DataConnection;
@@ -10,6 +11,8 @@ using ZB.MOM.WW.ScadaBridge.Commons.Messages.RemoteQuery;
using ZB.MOM.WW.ScadaBridge.Commons.Messages.Streaming;
using ZB.MOM.WW.ScadaBridge.Commons.Types;
using ZB.MOM.WW.ScadaBridge.Commons.Types.Alarms;
using ZB.MOM.WW.ScadaBridge.Commons.Types.Audit;
using ZB.MOM.WW.ScadaBridge.Commons.Types.Debugging;
using ZB.MOM.WW.ScadaBridge.Commons.Types.Enums;
namespace ZB.MOM.WW.ScadaBridge.Communication.Tests;
@@ -180,13 +183,13 @@ internal static class SiteCommandSamples
new ParkedOperationActionAck("corr-10", false, "handler missing"));
// ── Route commands ──
Add(new RouteToCallRequest("corr-12", "Site1.Pump1", "Start", Parameters(), T1, G1),
Add(new RouteToCallRequest("corr-12", "Site1.Pump1", "Start", Parameters(), T1, G1, Trace: true),
new RouteToCallRequest("corr-12", "Site1.Pump1", "Start", null, T2),
new RouteToCallRequest("corr-12", "Site1.Pump1", "Start", new Dictionary<string, object?>(), T1));
Add(new RouteToGetAttributesRequest("corr-13", "Site1.Pump1", ["Speed", "Flow"], T1, G1),
new RouteToGetAttributesRequest("corr-13", "Site1.Pump1", [], T2));
Add(new RouteToSetAttributesRequest(
"corr-14", "Site1.Pump1", new Dictionary<string, string> { ["Speed"] = "10" }, T1, G1),
"corr-14", "Site1.Pump1", new Dictionary<string, string> { ["Speed"] = "10" }, T1, G1, Trace: true),
new RouteToSetAttributesRequest("corr-14", "Site1.Pump1", new Dictionary<string, string>(), T2));
Add(new RouteToWaitForAttributeRequest(
"corr-15", "Site1.Pump1", "Speed", "10", TimeSpan.FromSeconds(30), T1, G1, true),
@@ -197,7 +200,7 @@ internal static class SiteCommandSamples
"corr-15", "Site1.Pump1", "Speed", "", TimeSpan.FromMinutes(1), T1));
// ── Route replies ──
Add(new RouteToCallResponse("corr-12", true, 17L, null, T1),
Add(new RouteToCallResponse("corr-12", true, 17L, null, T1, G1, TraceSamples()),
new RouteToCallResponse("corr-12", false, null, "script faulted", T2));
Add(new RouteToGetAttributesResponse("corr-13", Parameters(), true, null, T1),
new RouteToGetAttributesResponse("corr-13", new Dictionary<string, object?>(), false, "no instance", T2));
@@ -293,6 +296,30 @@ internal static class SiteCommandSamples
IsConfiguredPlaceholder = true
};
/// <summary>
/// Flow Debugger (design 2026-09-05 §5.4): two ScriptTrace rows for the maximal
/// RouteToCallResponse sample — a ScriptRun and an AttributeWrite with a populated Extra.
/// Fixed EventIds/timestamps so the golden is deterministic.
/// </summary>
private static IReadOnlyList<AuditEvent> TraceSamples() =>
[
ScadaBridgeAuditEventFactory.Create(
AuditChannel.ScriptTrace, AuditKind.ScriptRun, AuditStatus.Delivered,
eventId: Guid.Parse("aaaaaaaa-0000-0000-0000-000000000001"), occurredAtUtc: D1,
actor: "ScriptActor:DoWork", target: "DoWork", sourceNode: "node-a",
executionId: G1, parentExecutionId: Guid.Parse("bbbbbbbb-0000-0000-0000-000000000002"),
sourceSiteId: "site-1", sourceInstanceId: "Site1.Pump1", sourceScript: "ScriptActor:DoWork",
durationMs: 12, requestSummary: "{\"count\":3}", responseSummary: "7",
extra: new ScriptTraceExtra { TriggerKind = "call" }.ToJson()),
ScadaBridgeAuditEventFactory.Create(
AuditChannel.ScriptTrace, AuditKind.AttributeWrite, AuditStatus.Failed,
eventId: Guid.Parse("aaaaaaaa-0000-0000-0000-000000000003"), occurredAtUtc: D2,
actor: "ScriptActor:DoWork", target: "Speed", sourceNode: "node-a",
executionId: G1, sourceSiteId: "site-1", sourceInstanceId: "Site1.Pump1",
durationMs: 3, errorMessage: "device rejected",
extra: new ScriptTraceExtra { Writes = [new AttributeWriteDetail("Speed", "10", false, "device rejected")] }.ToJson()),
];
private static Dictionary<string, object?> Parameters() => new()
{
["count"] = 3,