Noderax Docs
Operations

Metrics, Events, Diagnostics, And Notifications

Use the operational observability surfaces and understand how live metric updates and workspace-scoped events move through the system.

Metrics, Events, Diagnostics, And Notifications

Noderax exposes several observability surfaces.

Metrics are ingested from the agent and exposed through both HTTP queries and realtime updates patched into the active web query cache.

Events capture workspace-scoped operational activity such as node state, task execution, install progress, and diagnostics signals.

Diagnostics views are API-backed operational surfaces used to inspect task flow, rollout behavior, update health, and runtime anomalies.

Notifications determine where operator-visible signals go, including mail and chat integrations configured by the workspace or platform admin.

Metrics

Metrics are ingested from the agent and made available through both the HTTP API and realtime updates to the web UI.

Recent HA fixes matter here:

  • metric list endpoints are distinct from Prometheus-style export paths
  • workspace-level realtime fanout covers live metric events
  • the web app patches its metrics query cache from websocket events

Events

Operational events capture node and task activity, status changes, and other workspace-scoped activity. In HA mode those events must survive cross-instance delivery, not just same-process emission.

Diagnostics

Diagnostics pages are used to investigate execution state, update flows, and runtime behavior. They are API-backed and usually workspace-aware.

Notifications

Notification delivery settings determine where operator signals go, including mail and chat integrations.

A healthy websocket handshake alone is not enough. Live metrics only feel correct when the API emits workspace-scoped events immediately and the web app consumes them into the active query cache.

What “live” depends on in HA mode

On this page