Feature Request
Add latency data visualization to the existing Metrics tab in SessionDetailPage and expand OverviewPage metric cards. The backend already tracks:
- State-change detection latency
- Channel delivery latency
- Permission response time
- Hook latency
Currently the backend returns this data via GET /v1/sessions/:id/latency and GET /v1/metrics, but the dashboard completely ignores it.
Why
- Latency metrics answer "how fast is Aegis responding?" — critical for debugging stalls
- The
GlobalMetrics type has 15+ fields that are fetched but only 4 are displayed
- No Zod schema exists for global metrics — the endpoint is unchecked
- Helps operators tune auto-approve thresholds and understand system health
Effort
M — New chart component, latency type definitions, poll/display logic, possibly a lightweight chart library
API Endpoints
GET /v1/sessions/:id/latencyGET /v1/metrics
Priority
P1 — Data is collected but invisible to operators
Feature Request
Add latency data visualization to the existing Metrics tab in SessionDetailPage and expand OverviewPage metric cards. The backend already tracks:
Currently the backend returns this data via
GET /v1/sessions/:id/latencyandGET /v1/metrics, but the dashboard completely ignores it.Why
GlobalMetricstype has 15+ fields that are fetched but only 4 are displayedEffort
M — New chart component, latency type definitions, poll/display logic, possibly a lightweight chart library
API Endpoints
GET /v1/sessions/:id/latencyGET /v1/metricsPriority
P1 — Data is collected but invisible to operators