Agents¶
The agents page shows every registered Noctuary Agent and the event stream it has sent.
Agent status¶
| Status | Meaning |
|---|---|
| Online | Last heartbeat within 2 minutes |
| Stale | Last heartbeat 2–10 minutes ago — check agent health |
| Offline | No heartbeat for over 10 minutes — agent may have stopped |
The agent sends a heartbeat automatically on every event ingest, so a healthy agent processing events will always show as online.
Event log¶
Click any agent to see its full event log — every ContextEvent the agent has sent, paginated with 10/20/50/100 events per page.
You can filter the event log by:
- Service — only show events from a specific service
- Vendor — only show events from a specific vendor (argocd, kubernetes, postgres)
- Event type — only show a specific event type (deploy, restart, saturation, etc.)
Event log columns¶
| Column | Description |
|---|---|
| Time | When the event was received |
| Service | Service name |
| Vendor | Which plugin detected the event |
| Type | Event type (deploy, restart, saturation, flag_change, etc.) |
| Entity | The specific resource affected (pod name, table name, app name) |
| Change | Old value and new value |
| Actor | Who or what caused the change (CI pipeline, user, etc.) |
| Confidence | Plugin confidence score (0.0–1.0) |
Multiple agents¶
You can run multiple agents — one per host, one per Kubernetes cluster, or one per environment. Each registers independently with its ingest API key and appears as a separate entry on this page.