perf(comms): alarms-only seed, capped buffers, at-least-once audit pull
This commit is contained in:
Joseph Doherty
@@ -66,18 +66,23 @@ Both central and site clusters. Each side has communication actors that handle m
|
||||
- Central sends an unsubscribe request over `SiteCommandService` (gRPC command/control) when the debug session ends. The gRPC stream is cancelled. The site's `StreamRelayActor` is stopped and the SiteStreamManager subscription is removed.
|
||||
- The stream is session-based and temporary.
|
||||
- **Accepted limitation (central-side session locality):** debug sessions are process-local state on the central node hosting the `DebugStreamBridgeActor`. A central **restart or failover drops all active sessions** with no `OnStreamTerminated` signal to the operator — the engineer simply re-establishes the debug session from the UI. Debug streaming is an interactive, transient diagnostic aid, so this is an accepted trade-off rather than a durability gap (arch review 02, U7).
|
||||
- **Backpressure is lossy-by-design and now observable:** the per-session `Channel<SiteStreamEvent>` (bounded 1000, `DropOldest`) silently evicts the oldest event when a slow consumer falls behind. The eviction is counted via the channel's `itemDropped` callback and logged at Warning (first eviction, then every 500th) so real event loss on a debug stream is visible instead of silent.
|
||||
- **Backpressure is lossy-by-design and now observable:** the per-session `Channel<SiteStreamEvent>` (bounded `GrpcInstanceStreamChannelCapacity`, default 1000, `DropOldest`) silently evicts the oldest event when a slow consumer falls behind. The eviction is counted via the channel's `itemDropped` callback, logged at Warning (first eviction, then every 500th) and exported as `scadabridge.site.stream.events_dropped` (tagged `stream=instance`), so real event loss on a debug stream is visible instead of silent. The **site-wide alarm feed has its own, larger channel** — see §6.1.
|
||||
- **Bridge-session hardening (WP2.3):**
|
||||
- The pre-snapshot buffer is **bounded (20 000 events, drop-oldest)** and its evictions counted (`scadabridge.central.debug_view.presnapshot_dropped`). It was previously unbounded, so a session whose snapshot never arrived grew without limit on the CENTRAL node. Dropping the oldest is correct here: the snapshot that ends the buffering phase is authoritative for anything that old.
|
||||
- A **hard snapshot deadline** (`DebugStreamBridgeActor.SnapshotTimeout`, 60 s) fails the session if no `DebugViewSnapshot` arrives. Nothing else ended a session wedged in the buffering phase — a lost site reply raises no gRPC error.
|
||||
- **Stream events no longer influence the orphan receive timeout.** The gRPC callback wraps each event in an envelope marked `INotInfluenceReceiveTimeout`, so a busy site can no longer keep an abandoned session alive forever by feeding it events. The 5-minute timeout now measures session/consumer liveness, which is what it was for.
|
||||
|
||||
### 6.1 Aggregated Live Alarm Stream (Site → Central)
|
||||
|
||||
Delivered 2026-07-10 (`docs/plans/2026-07-10-aggregated-live-alarm-stream-plan.md`) for the operator Alarm Summary page (Central UI #9). Where §6 streams **one instance** for the interactive Debug View, this feeds a **transient, per-site central live alarm cache** so the Alarm Summary reflects alarm transitions in near-real-time instead of re-polling the whole site every 15s.
|
||||
|
||||
- **Site-wide, alarm-only stream**: a single `SubscribeSite` gRPC stream per site carries `AlarmStateChanged` events for **all** instances (attribute events are dropped — the summary never shows them, and attributes are far higher-volume). On the site this reuses the existing `SiteStreamManager` broadcast hub via `SiteStreamManager.SubscribeSiteAlarms(...)` (same wiring as the per-instance `Subscribe`, minus the `InstanceUniqueName` filter and filtered to alarms). `StreamRelayActor` is reused unchanged, so placeholder rows (`is_configured_placeholder`) are still dropped at the relay.
|
||||
- **Central live cache** (`ISiteAlarmLiveCache`, singleton `SiteAlarmLiveCacheService`): a DI singleton on the active central node. For each site with ≥1 active viewer it runs ONE shared, **reference-counted** per-site aggregator (`SiteAlarmAggregatorActor`); the first `Subscribe(siteId, onChanged)` starts it, the last subscriber leaving stops it after a short **linger** to avoid re-seed thrash. `GetCurrentAlarms(siteId)` returns the current immutable snapshot; `IsLive(siteId)` reports whether the aggregator has seeded and published at least once.
|
||||
- **Central live cache** (`ISiteAlarmLiveCache`, singleton `SiteAlarmLiveCacheService`): a DI singleton on the active central node. For each site with ≥1 active viewer it runs ONE shared, **reference-counted** per-site aggregator (`SiteAlarmAggregatorActor`); the first `Subscribe(siteId, onChanged)` starts it, the last subscriber leaving stops it after a short **linger** to avoid re-seed thrash. `GetCurrentAlarms(siteId)` returns the current immutable snapshot; `IsLive(siteId)` reports whether the aggregator has seeded **and its site-wide stream is currently up**. Liveness rides every publish from the aggregator, so a stream that faults or ends gracefully drops `IsLive` immediately rather than leaving the page grafting a freezing snapshot over fresh poll data until the next reconcile (WP2.3).
|
||||
- **Seed-then-stream** (copied from `DebugStreamBridgeActor` ordering): open the `SubscribeSite` stream first (buffer live deltas), run the snapshot fan-out once via the existing `DebugViewSnapshot` path (bounded by `LiveAlarmCacheSeedConcurrency`), flush the buffer with **dedup by `(InstanceUniqueName, AlarmName, SourceReference)`**, then live pass-through. Placeholders are seeded from the snapshot and never expected on the live stream.
|
||||
- **Failover & drift**: NodeA↔NodeB reconnect (client via `SiteStreamGrpcClient.SubscribeSiteAsync`, factory endpoint failover) re-seeds rather than serving stale; a periodic **reconcile snapshot** (default 60s) corrects any missed enable/disable so the cache can't drift indefinitely. `[PERM]` (`docs/plans/2026-05-29-native-alarms-design.md`): the cache is **purely in-memory** — no EF entity/table/migration, no persisted central alarm store — so a new active node simply re-seeds from scratch.
|
||||
- **Options** (`Communication` section, `CommunicationOptions`; eagerly validated by `CommunicationOptionsValidator` / `ValidateOnStart`): `LiveAlarmCacheLinger` (default 30s), `LiveAlarmCacheReconcileInterval` (default 60s), `LiveAlarmCacheSeedConcurrency` (default 8), `LiveAlarmCacheMaxSubscribersPerSite` (default 200), `LiveAlarmCachePublishCoalesce` (default 250ms; `0` = publish per delta — legacy — batches an alarm storm into one snapshot copy + one viewer fan-out per window; arch review 02 round 2, N6).
|
||||
- **Telemetry** (`ScadaBridgeTelemetry` meter): observable gauge `scadabridge.site.alarm_cache.aggregators.active` (running per-site aggregators) and counter `scadabridge.site.alarm_cache.reconnects` (site-wide stream reconnects — a NodeA↔NodeB flip or reconcile-driven reopen; a sustained climb signals a flapping site link).
|
||||
- **Alarms-only seed (WP2.3)**: the seed/reconcile fan-out sets `DebugSnapshotRequest.AlarmsOnly` (wire: `DebugSnapshotRequestDto.alarms_only`, field 3, additive), so the site builds and ships only the alarm half of the snapshot. The fan-out discarded every attribute row anyway, and an instance's attribute surface dwarfs its alarm set. A pre-WP2.3 site ignores the flag and returns the full snapshot, which reads identically.
|
||||
- **Failover & drift**: a re-seed runs **once per successful (re)connect**, not once per reconnect ATTEMPT — the connect signal is `SiteStreamGrpcClient.SubscribeSiteAsync`'s `onConnected` callback, raised when the site's response headers arrive (the site flushes them as soon as its relay is attached, so no event can be missed after it). Fanning a whole-site snapshot out per retry meant N snapshots against a site that was, by definition of the retry, unreachable. A periodic **reconcile snapshot** (default 60s, **jittered** by `LiveAlarmCacheReconcileJitterFraction` so aggregators started together do not stampede one boundary) remains the drift backstop, but it is **skipped when a fan-out already ran in that window** and **publishes only when the snapshot actually changed the cache** (a diff, not an unconditional viewer fan-out). Staleness stays bounded at two intervals: the skip consumes its flag, so the next tick always fans out. A fan-out that fails as a whole now retries on its own **backoff** timer (`reconnectDelay` doubling, capped at 8× the reconcile interval) instead of waiting a full interval. `[PERM]` (`docs/plans/2026-05-29-native-alarms-design.md`): the cache is **purely in-memory** — no EF entity/table/migration, no persisted central alarm store — so a new active node simply re-seeds from scratch.
|
||||
- **Options** (`Communication` section, `CommunicationOptions`; eagerly validated by `CommunicationOptionsValidator` / `ValidateOnStart`): `LiveAlarmCacheLinger` (default 30s), `LiveAlarmCacheReconcileInterval` (default 60s), `LiveAlarmCacheSeedConcurrency` (default 8), `LiveAlarmCacheMaxSubscribersPerSite` (default 200), `LiveAlarmCachePublishCoalesce` (default 250ms; `0` = publish per delta — legacy — batches an alarm storm into one snapshot copy + one viewer fan-out per window; arch review 02 round 2, N6), `LiveAlarmCacheReconcileJitterFraction` (default 0.2 = up to +20% per tick; `0` disables). Stream channel sizing lives alongside the other gRPC limits: `GrpcInstanceStreamChannelCapacity` (default 1000) and `GrpcSiteAlarmStreamChannelCapacity` (default 20 000).
|
||||
- **Telemetry** (`ScadaBridgeTelemetry` meter): observable gauge `scadabridge.site.alarm_cache.aggregators.active` (running per-site aggregators) and counter `scadabridge.site.alarm_cache.reconnects` (site-wide stream reconnects — a NodeA↔NodeB flip or reconcile-driven reopen; a sustained climb signals a flapping site link), and counter `scadabridge.site.alarm_cache.buffer_dropped` (deltas evicted from the aggregator's **bounded** 20 000-entry pre-seed buffer, drop-oldest — non-zero means a fan-out ran long enough for the delta storm behind it to exceed the cap; the fan-out's snapshot is authoritative for the evicted rows).
|
||||
- **Accepted limitations (arch review 02 round 2, N8):**
|
||||
- **Standby-node aggregators**: the live cache is per-node and `SetActorSystem` is wired on every central node, so browsing the standby node directly (diagnostic ports, e.g. 9002) starts a second, fully-functional read-only aggregator + `SubscribeSite` stream there. Accepted — not gated: the aggregator is read-only, bounded (one stream/site, viewer-capped), and torn down by the viewer linger; gating `SetActorSystem` behind the active check would break the diagnostic-browse path and buy nothing. The `[PERM]` "lives only on the active central node" claim is corrected to "per-node; in routine operation only the active node hosts viewers".
|
||||
- **Site deleted while an Alarm Summary viewer is open**: the viewer's aggregator reconciles to an empty snapshot (the deleted site's instances vanish from the fan-out) until its viewers leave, then the linger stop reaps it; the site's cached gRPC channels are disposed at deletion via `SiteStreamGrpcClientFactory.RemoveSiteAsync` (see `ManagementActor.HandleDeleteSite`, arch review 02 round 2, N8).
|
||||
@@ -101,12 +106,12 @@ The streaming protocol is defined in `sitestream.proto` (`src/ZB.MOM.WW.ScadaBri
|
||||
|
||||
- **Service**: `SiteStreamService` — hosted on each site node by `SiteStreamGrpcServer` — exposes six RPCs. Two are real-time **server-streaming** subscriptions (`SubscribeInstance`, `SubscribeSite`); the other four are **unary request/response** calls added by the Audit Log (#23) and Site Call Audit (#22) components. These are distinct from the command/control gRPC services (`CentralControlService` on central, `SiteCommandService` on the site) — `SiteStreamService` is the streaming + audit-pull surface, not the command/control channel:
|
||||
- `SubscribeInstance(InstanceStreamRequest) returns (stream SiteStreamEvent)` — the per-instance real-time debug stream (§6).
|
||||
- `SubscribeSite(SiteStreamRequest) returns (stream SiteStreamEvent)` — the **site-wide, alarm-only** aggregated stream (§6.1) added for the operator Alarm Summary live cache. Additive (new RPC + new `SiteStreamRequest { correlation_id }` message; no field renumbering). The server handler mirrors `SubscribeInstance` — same bounded `Channel(1000, DropOldest)`, `GrpcMaxConcurrentStreams`/`GrpcMaxStreamLifetime` limits, `SiteConnectionOpened/Closed` telemetry, and `StreamRelayActor` mapping — but subscribes via `SiteStreamManager.SubscribeSiteAlarms` (all instances, `AlarmStateChanged` only; attribute events dropped, no per-instance filter).
|
||||
- `SubscribeSite(SiteStreamRequest) returns (stream SiteStreamEvent)` — the **site-wide, alarm-only** aggregated stream (§6.1) added for the operator Alarm Summary live cache. Additive (new RPC + new `SiteStreamRequest { correlation_id }` message; no field renumbering). The server handler mirrors `SubscribeInstance` — but with **its own, larger send channel** (`GrpcSiteAlarmStreamChannelCapacity`, default 20 000, `DropOldest`; WP2.3): sharing the Debug View's 1000-slot channel meant an alarm burst during a WAN stall silently evicted operator-visible transitions to make room for diagnostics traffic. Drops are tagged `stream=site-alarms` on `scadabridge.site.stream.events_dropped`. Otherwise identical — same `GrpcMaxConcurrentStreams`/`GrpcMaxStreamLifetime` limits, `SiteConnectionOpened/Closed` telemetry, and `StreamRelayActor` mapping — but subscribes via `SiteStreamManager.SubscribeSiteAlarms` (all instances, `AlarmStateChanged` only; attribute events dropped, no per-instance filter).
|
||||
- `IngestAuditEvents(AuditEventBatch) returns (IngestAck)` — a legacy central-side ingest surface on the *site*-hosted service; it is **dead in the shipped topology** because no site ever dials a site for ingest. The production audit-telemetry *push* path is site→central gRPC to `CentralControlService`, which routes the batch to the central `AuditLogIngestActor` proxy and returns the accepted `EventId`s.
|
||||
- `IngestCachedTelemetry(CachedTelemetryBatch) returns (IngestAck)` — ingest receiving surface for the combined cached-call telemetry packet (audit row + `SiteCalls` operational upsert written in one transaction).
|
||||
- `PullAuditEvents(PullAuditEventsRequest) returns (PullAuditEventsResponse)` — central→site **reconciliation pull** for the Audit Log self-heal feed; the site serves `Pending`/`Forwarded` rows from its `ISiteAuditQueue`.
|
||||
- `PullAuditEvents(PullAuditEventsRequest) returns (PullAuditEventsResponse)` — central→site **reconciliation pull** for the Audit Log self-heal feed; the site serves `Pending`/`Forwarded` rows from its `ISiteAuditQueue`. **At-least-once (WP2.3):** rows are NOT retired when served. The site retires (`Reconciled`) everything at or before the cursor of the **next** pull, because that cursor is central's only proof of receipt — a fault between the response leaving the site and central committing it re-serves the rows instead of losing them (central dedups on `EventId`). The request carries an additive composite-keyset `after_id` (field 3) mirroring `PullSiteCallsRequest`; with it the read is a strict `(OccurredAtUtc, EventId)` keyset and the retirement is exact, without it the legacy inclusive `>=` read applies and only rows strictly older than the cursor instant are provably received.
|
||||
- `PullSiteCalls(PullSiteCallsRequest) returns (PullSiteCallsResponse)` — central→site reconciliation pull for the Site Call Audit (#22) self-heal feed; the site serves operation-tracking rows changed since a cursor from its `IOperationTrackingStore`. A separate RPC from `PullAuditEvents` because the tracking store is the operational source of truth, distinct from the site audit queue.
|
||||
- **Messages**: `InstanceStreamRequest` (correlation_id, instance_unique_name), `SiteStreamRequest` (correlation_id only — no instance name; drives the site-wide `SubscribeSite` stream), `SiteStreamEvent` (correlation_id, oneof event: `AttributeValueUpdate`, `AlarmStateUpdate`); `AuditEventDto`/`AuditEventBatch`/`IngestAck` for ingest; `CachedTelemetryPacket`/`CachedTelemetryBatch` (each packet pairing an `AuditEventDto` with a `SiteCallOperationalDto`); `PullAuditEventsRequest`/`PullAuditEventsResponse` and `PullSiteCallsRequest`/`PullSiteCallsResponse` (each request carries `since_utc` + `batch_size`; each response carries `more_available` to signal a saturated batch).
|
||||
- **Messages**: `InstanceStreamRequest` (correlation_id, instance_unique_name), `SiteStreamRequest` (correlation_id only — no instance name; drives the site-wide `SubscribeSite` stream), `SiteStreamEvent` (correlation_id, oneof event: `AttributeValueUpdate`, `AlarmStateUpdate`); `AuditEventDto`/`AuditEventBatch`/`IngestAck` for ingest; `CachedTelemetryPacket`/`CachedTelemetryBatch` (each packet pairing an `AuditEventDto` with a `SiteCallOperationalDto`); `PullAuditEventsRequest`/`PullAuditEventsResponse` and `PullSiteCallsRequest`/`PullSiteCallsResponse` (each request carries `since_utc` + `batch_size` + an optional composite-keyset `after_id`; each response carries `more_available` to signal a saturated batch).
|
||||
- The `oneof event` pattern is extensible — future event types (health metrics, connection state changes) are added as new fields without breaking existing consumers.
|
||||
- Proto field numbers are never reused; new RPCs and message fields are appended additively. Old clients ignore unknown `oneof` variants.
|
||||
|
||||
@@ -162,7 +167,7 @@ read the traffic". TLS is follow-on hardening and does not change this design.
|
||||
- **Placeholder rows are dropped at the relay**: `is_configured_placeholder` (field 23) is a **Debug View snapshot-only** concept emitted by `InstanceActor.BuildAlarmStatesSnapshot` for quiet bindings — it is never a real alarm transition (its timestamp may be `DateTimeOffset.MinValue`, the Protobuf `Timestamp` lower boundary). `StreamRelayActor.HandleAlarmStateChanged` therefore returns early — **never relaying a placeholder row to the live gRPC stream** — so field 23 is always `false` on the live stream and only ever carries `true` in the snapshot path.
|
||||
- **Client-side mapping (`SiteStreamGrpcClient.ConvertToDomainEvent`)**: reconstructs the domain `AlarmStateChanged` from the proto — `Kind` is parsed via `ParseAlarmKind`, the `Condition` is rebuilt with `severity` taken from the existing wire `priority`, and native metadata is repopulated from fields 8–23 (`native_source_canonical_name` → `NativeSourceCanonicalName`, `is_configured_placeholder` → `IsConfiguredPlaceholder`) — so central-side consumers receive the same domain event the site emitted.
|
||||
|
||||
> **Regeneration is manual (macOS-only).** `sitestream.proto` is **not** auto-compiled: the `<Protobuf>` include is commented out in the `.csproj`, and the generated C# is **vendored** under `SiteStreamGrpc/`. To regenerate after editing the proto: toggle the `<Protobuf>` include on, build so `Grpc.Tools` regenerates the C#, copy the generated files into `SiteStreamGrpc/`, then re-comment the include. Adding `AlarmStateUpdate` fields 8–23 and the four unary RPCs (`IngestAuditEvents`, `IngestCachedTelemetry`, `PullAuditEvents`, `PullSiteCalls`) plus their message types followed this process.
|
||||
> **Regeneration is manual (macOS-only).** `sitestream.proto` is **not** auto-compiled: the `<Protobuf>` include is commented out in the `.csproj`, and the generated C# is **vendored** under `SiteStreamGrpc/`. To regenerate after editing the proto: toggle the `<Protobuf>` include on, build so `Grpc.Tools` regenerates the C#, copy the generated files into `SiteStreamGrpc/`, then re-comment the include. Adding `AlarmStateUpdate` fields 8–23 and the four unary RPCs (`IngestAuditEvents`, `IngestCachedTelemetry`, `PullAuditEvents`, `PullSiteCalls`) plus their message types followed this process, as did WP2.3's two additive fields — `PullAuditEventsRequest.after_id` (field 3, `sitestream.proto`) and `DebugSnapshotRequestDto.alarms_only` (field 3, `site_command.proto`). The same applies to `site_command.proto`/`SiteCommandGrpc/` and `central_control.proto`/`CentralControlGrpc/`; `docker/regen-proto.sh [sitestream|centralcontrol|sitecommand|all]` automates the toggle-build-copy-untoggle and always restores the csproj.
|
||||
|
||||
#### gRPC Connection Keepalive
|
||||
|
||||
|
||||
Reference in New Issue
Block a user