feat(security): re-query LDAP group membership for active sessions (#46)

Auth 0.5.0 (all five packages) adds the passwordless ILdapGroupLookup. A new
singleton DirectoryGroupRefresher asks it for a session's current groups once
the new zb:lastgrouprefresh anchor (stamped at login) is older than
ScadaBridge:Security:LdapGroupRefresh:IntervalMinutes (default 15, Enabled
default true, CacheSeconds 60 per-node answer reuse). CookieSessionValidator
replaces the stored groups and re-maps roles; this reaches HTTP and live
circuits (once per interval, independent of the 10 s circuit refresh).

Failure policy: directory unavailable keeps the stored groups (warn, all paths
incl. fail-closed mutation checks); user not found / disabled / ambiguous /
no groups ends the session. Inert under BindMode=User (NotSupported),
LDAP disabled, or DisableLogin.

Auth 0.4.4 ApiKeyFailure.StoreUnavailable now answers 503 SERVICE_UNAVAILABLE
on the inbound API instead of a misleading 401.

Docs: Component-Security.md, components/Security.md, Component-Host.md,
register row 23 closed.

Claude-Session: https://claude.ai/code/session_01GT67ACmpMR7b9LGRPLijMi
This commit is contained in:
Joseph Doherty
2026-11-12 20:31:51 -05:00
parent 139a43fb25
commit 9b0720af9c
16 changed files with 919 additions and 47 deletions
+7 -4
View File
@@ -21,11 +21,11 @@ The component registers:
Authentication uses a direct username/password bind against the LDAP/AD server via the shared `ILdapAuthService` (`ZB.MOM.WW.Auth.Ldap`). The flow is: service-account bind → search for the user entry by username → user-credential bind → group-membership query. The app never caches credentials locally. LDAPS (port 636) or StartTLS is required for production; the `AllowInsecure` flag in `LdapOptions` gates unencrypted use to explicitly opted-in deployments (local dev only). No Kerberos/NTLM path exists.
LDAP failure behavior is fail-closed at the login boundary and fail-open at the session boundary: a new login fails immediately if the directory is unreachable; an active cookie session continues with its current claims (its mid-session role refresh is DB-only and never calls LDAP), and a CLI holding a bearer token keeps working until the token expires. This avoids disrupting engineers mid-work during a brief directory outage. Group-membership changes are picked up at the next login.
LDAP failure behavior is fail-closed at the login boundary and fail-open at the session boundary: a new login fails immediately if the directory is unreachable; an active cookie session continues with its current claims (its periodic directory group re-query keeps the stored groups when the directory cannot answer), and a CLI holding a bearer token keeps working until the token expires. This avoids disrupting engineers mid-work during a brief directory outage. Group-membership changes reach an active session within `LdapGroupRefresh:IntervalMinutes` (default 15, ScadaBridge#46); a deleted or disabled account ends the session.
### Sessions: cookie (UI) and bearer JWT (CLI/API)
**Interactive UI.** `POST /auth/login` signs the user in with ASP.NET Core cookie authentication. The HttpOnly/Secure cookie carries a data-protected claims ticket (no embedded JWT) built by the shared `SessionClaimBuilder`: display name, username, roles, permitted site IDs (site-scoped Deployer), the raw LDAP groups and the idle / role-refresh anchors. The cookie is the transport because Blazor Server's SignalR circuits do not carry `Authorization` headers on reconnect. Both central nodes share the Data Protection key ring, so either node validates a cookie the other issued. The cookie's `ExpireTimeSpan` is `IdleTimeoutMinutes` (default 30) with sliding expiration; passive traffic does not renew it. Once the role-refresh anchor is older than `RoleRefreshThresholdMinutes` (default 15) the validation hook re-runs the DB-backed `RoleMapper` on the stored groups (no LDAP call).
**Interactive UI.** `POST /auth/login` signs the user in with ASP.NET Core cookie authentication. The HttpOnly/Secure cookie carries a data-protected claims ticket (no embedded JWT) built by the shared `SessionClaimBuilder`: display name, username, roles, permitted site IDs (site-scoped Deployer), the raw LDAP groups and the idle / role-refresh anchors. The cookie is the transport because Blazor Server's SignalR circuits do not carry `Authorization` headers on reconnect. Both central nodes share the Data Protection key ring, so either node validates a cookie the other issued. The cookie's `ExpireTimeSpan` is `IdleTimeoutMinutes` (default 30) with sliding expiration; passive traffic does not renew it. Once the role-refresh anchor is older than `RoleRefreshThresholdMinutes` (default 15) the validation hook re-runs the DB-backed `RoleMapper` on the stored groups. Separately, once the `zb:lastgrouprefresh` anchor is older than `LdapGroupRefresh:IntervalMinutes` (default 15), the groups themselves are re-read from the directory with the LDAP service account (`DirectoryGroupRefresher` over Auth `ILdapGroupLookup`, ScadaBridge#46) before the re-map.
**CLI / API.** `POST /auth/token` performs one LDAP bind per invocation and returns a short-lived HS256 JWT (`JwtExpiryMinutes`, default 15) signed with `JwtSigningKey`, shared by both central nodes; the response advertises `management_bearer: true`. `ManagementAuthenticator` accepts `Authorization: Bearer` alongside HTTP Basic on `POST /management`, the audit REST endpoints and the debug-stream hub. Roles inside a valid token are authoritative until it expires, so a revocation reaches a bearer caller within `JwtExpiryMinutes`; rotating `JwtSigningKey` invalidates every token. Under `DisableLogin`, `/auth/token` answers 404. Full spec: `docs/requirements/Component-Security.md`.
@@ -138,7 +138,7 @@ Login flow (Central UI `/auth/login` and `/auth/token`):
2. On success, call `RoleMapper.MapGroupsToRolesAsync(ldapGroups)` to resolve roles and site scope.
3. `/auth/login`: build the claims via `SessionClaimBuilder` and call the ASP.NET cookie `SignInAsync`. `/auth/token`: call `JwtTokenService.GenerateToken(displayName, username, roles, permittedSiteIds)` and return the bearer JWT.
On each subsequent cookie request the validation hook enforces the idle timeout and, past the role-refresh threshold, re-maps the stored groups against the DB. Bearer requests validate the token with `JwtTokenService.ValidateToken` and perform no LDAP bind or role re-map.
On each subsequent cookie request the validation hook enforces the idle timeout and, past the role-refresh threshold, re-maps the stored groups against the DB (re-reading them from the directory first when the group re-query is due). Bearer requests validate the token with `JwtTokenService.ValidateToken` and perform no LDAP bind or role re-map.
Authorization gates use the named policies:
@@ -162,7 +162,10 @@ if (!AuthorizationPolicies.OperationalAuditRoles.Contains(userRole))
| `JwtSigningKey` | *(required)* | Symmetric HMAC-SHA256 signing key. Must be at least 32 bytes (256 bits); `StartupValidator` rejects a shorter key on Central. |
| `JwtExpiryMinutes` | `15` | Lifetime of the bearer JWT issued by `/auth/token`. |
| `IdleTimeoutMinutes` | `30` | Cookie session idle timeout; set as the cookie `ExpireTimeSpan`. |
| `RoleRefreshThresholdMinutes` | `15` | Age of the role-refresh anchor after which the cookie's roles are re-mapped from the DB (LDAP-free). |
| `RoleRefreshThresholdMinutes` | `15` | Age of the role-refresh anchor after which the cookie's roles are re-mapped from the DB. |
| `LdapGroupRefresh:Enabled` | `true` | Mid-session re-query of the user's directory groups (ScadaBridge#46). Inert when LDAP has no service account (`BindMode = User`) or under `DisableLogin`. |
| `LdapGroupRefresh:IntervalMinutes` | `15` | Age of the `zb:lastgrouprefresh` anchor after which the groups are re-read from the directory (≥ 1). |
| `LdapGroupRefresh:CacheSeconds` | `60` | Per-node reuse of one user's directory answer (0 disables). |
| `RoleMappingCacheSeconds` | `30` | Process-wide `RoleMappingCache` TTL (`0` disables); invalidated after mapping/scope-rule edits on the local node and, over the central cache-invalidation topic, on the peer central node (#100). |
| `RequireHttpsCookie` | `true` | When `true`, the cookie is `Secure`-only (HTTPS required). Set `false` for HTTP-only dev deployments; a warning is logged at startup. |
| `CookieName` | `ZB.MOM.WW.ScadaBridge.Auth` | Authentication cookie name. Override per deployment when two ScadaBridge stacks share a hostname — browsers scope cookies by host+path but not by port. |