Lightweight containerized prosody tooling + moved auth scripts + xmpp reconnect/auth stabilization
This commit is contained in:
27
artifacts/plans/05-adapter-resilience-supervisor.md
Normal file
27
artifacts/plans/05-adapter-resilience-supervisor.md
Normal file
@@ -0,0 +1,27 @@
|
||||
# Feature Plan: Adapter Resilience Supervisor
|
||||
|
||||
## Goal
|
||||
Make adapters self-healing and observable under disconnects, API drift, and transient faults.
|
||||
|
||||
## Why This Fits GIA
|
||||
- Bridge reliability is a product requirement for multi-network ops.
|
||||
|
||||
## Scope
|
||||
- Health probes per adapter.
|
||||
- Reconnect/backoff supervisor.
|
||||
- Circuit-breaker for repeated failure classes.
|
||||
|
||||
## Implementation
|
||||
1. Add adapter health state model (healthy/degraded/down).
|
||||
2. Add watchdog jobs for Signal/WhatsApp/Instagram/XMPP.
|
||||
3. Implement exponential backoff + jitter reconnect policies.
|
||||
4. Emit structured adapter health events + alerts.
|
||||
5. Add status surface in settings/services pages.
|
||||
|
||||
## Acceptance Criteria
|
||||
- Adapter restarts automatically on transient failures.
|
||||
- Repeated failures degrade state and stop spam retries.
|
||||
- Operators can view current health + last error + retry schedule.
|
||||
|
||||
## Out of Scope
|
||||
- Full protocol hot-reload without process restart.
|
||||
Reference in New Issue
Block a user