root/fusion
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

feat(FN-3960): add mesh replay outage contracts and runtime hooks

Browse Source 
- Add core outage schema and mesh replay contract types, with CentralCore/CentralDB support for queue and snapshot persistence
- Expose runtime replay hooks in engine health monitoring and peer exchange paths to drive outage recovery flows
- Expand protocol and architecture docs for shared mesh replay behavior and finalized contract expectations
- Add regression tests across core and engine for queue sequencing, snapshot contracts, and replay integration

Fusion-Task-Id: FN-3960
This commit is contained in:
Fusion
2026-05-10 19:11:31 -07:00
committed by gsxdsm
parent 7d67dc3597
commit a5a794d12e
14 changed files with 717 additions and 15 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
docs/architecture.md
View File

@@ -621,9 +621,9 @@ Implemented in `agent-heartbeat.ts`:
- `MeshLeaseManager` (`mesh-lease-manager.ts`) — canonical abandoned-lease detection + recovery path
### Outage ownership boundaries (degraded reads + queued write replay)
- `CentralCore` owns durable outage state in central persistence (`meshSharedSnapshots` + `meshWriteQueue`) and computes canonical degraded-read metadata (`MeshDegradedReadState`) for API consumers.
- `PeerExchangeService` owns retryability classification for sync/apply failures, queue insertion for retryable failures, and replay execution after successful peer sync using queue ordering `(createdAt ASC, id ASC)`.
- `NodeHealthMonitor` provides liveness transitions as replay hints only; `online` is a trigger to attempt replay, not proof that replay succeeded.
- `CentralCore` owns durable outage state in central persistence (`meshSharedSnapshots` + `meshWriteQueue`) and exposes stable assertion methods: `recordMeshSnapshot`, `getLatestMeshSnapshot`, `enqueueMeshWrite`, `listPendingMeshWrites`, `markMeshWriteReplayStarted`, `markMeshWriteApplied`, `markMeshWriteFailed`, and `getMeshDegradedReadState`.
- `PeerExchangeService` owns retryability classification for sync/apply failures, queue insertion for retryable failures, replay execution (`replayPendingWritesForNode(targetNodeId)`), and observable sync results (`queuedWriteId`, `replaySummary`) for partition/replay assertions.
- `NodeHealthMonitor` provides liveness transitions as replay hints only via deterministic recovery callback `onNodeRecovered(nodeId, previousStatus)`; `online` is a trigger to attempt replay, not proof that replay succeeded.
- Dashboard mesh routes (`register-mesh-routes.ts`) preserve `GET /api/mesh/state` array shape and attach per-node degraded `readState` metadata so stale fallback data is explicit during partitions.
### Mesh task lease ownership and recovery