THE FIELD GUIDE / OPERATIONS

Deployment health & worker readiness

See expected components, missing nodes and the difference between alive, ready and current.

Mr Patchy documentation · 2 min read

Inventory what should be there

The deployment dashboard tracks expected and registered controllers, workers, integration nodes and managed endpoint agents, so stopped or missing components remain visible. Show identity, role, machine/location, version/image, last heartbeat, available uptime, health reason and current work. Endpoint agents retain their actual Docker/systemd deployment type.

Separate liveness, readiness and freshness

A running process is not necessarily admitted to claim work, and a ready controller is not proof of fresh endpoint telemetry. Identity/key agreement, schema structure, receipt protocol, version compatibility, required state and permissions contribute to role readiness. Healthy, degraded, offline and unknown states have explicit heartbeat/freshness thresholds and explained failures.

Inspect worker capacity and drain

Configured concurrency, queued/active jobs, last contact and drain state link to relevant jobs, configuration and diagnostics. Expected-worker inventory distinguishes an unregistered/missing worker from an accepted worker with stale heartbeats. Controlled diagnostics explain startup admission without exposing credentials or raw endpoint evidence.

Include integrations and available resources

GCP integration readiness and failures appear alongside workers and controllers. Available resource metrics and recent failure summaries retain explicit unavailable states. Authenticated reporting and bounded read-only checks collect supported metadata without granting the web app unrestricted Docker-daemon control.

Turn health into attention

Meaningful node, agent, service and coverage problems feed persistent alerts and live notifications. A connected event stream only establishes stream connectivity; inspect component readiness and collection timestamps separately. Plan upgrades and recovery from the actual affected role and shared-state dependencies.

Go deeper

Open the application reference for capability scope and technical details.

Open Open SECURITY_IMPLEMENTATION.md on GitHub ↗

Actual application interface with illustrative documentation data.