28 lines
969 B
Markdown
28 lines
969 B
Markdown
# Feature Plan: End-to-End Observability and Traceability
|
|
|
|
## Goal
|
|
Provide trace-level visibility from ingress transport event to UI delivery/ack.
|
|
|
|
## Why This Fits GIA
|
|
- Multi-hop messaging systems require correlation IDs to debug reliably.
|
|
|
|
## Scope
|
|
- Global trace IDs for message lifecycle.
|
|
- Structured logs and timeline diagnostics view.
|
|
- Basic metrics and SLA dashboards.
|
|
|
|
## Implementation
|
|
1. Inject `trace_id` at ingress/send initiation.
|
|
2. Propagate through router, persistence, websocket, command/task flows.
|
|
3. Standardize structured log schema across services.
|
|
4. Add timeline diagnostics page by trace ID and session.
|
|
5. Add core metrics: ingress latency, send latency, drop rate, retry counts.
|
|
|
|
## Acceptance Criteria
|
|
- One trace ID can reconstruct full message path.
|
|
- At least 95% of critical paths emit structured trace logs.
|
|
- Operators can isolate bottleneck stage in under 2 minutes.
|
|
|
|
## Out of Scope
|
|
- Full distributed tracing vendor integration.
|