Skip to content

Engine Operations

This page covers routine day-2 operations for MeshFlows Engine.

Health Checks

kubectl get pods -n meshflows
curl http://gateway:8080/v1/status
curl http://orchestrator:8083/readyz
# Unknown public edge route should return 404 and not invoke orchestrator.
curl -i -X POST http://gateway:8080/v1/api/unknown -H "Content-Type: application/json" -d '{"xml":"<r/>"}'

Runtime Changes

  • Update workflows/connections/policies through your deployment process.
  • Update edge route configuration (EDGE_ROUTE_CONFIG_JSON) when adding public alias routes.
  • Trigger orchestrator reload when needed.
  • Verify new versions with trace and trigger-event APIs.

Incident Triage

  1. Check readiness and service status.
  2. Inspect failed traces and trigger events.
  3. Inspect queue depth and broker health.
  4. Check recent config/secret changes.

Scaling

  • Scale stateless components (gateway/orchestrator) based on load.
  • Watch downstream bottlenecks (transformers, storage, broker).
  • Validate idempotency/throttle behavior during burst traffic.

Maintenance

  • Keep image versions and dependencies current.
  • Rotate credentials and tokens.
  • Archive or expire old trace data per retention policy.

Remote cluster access

For SSH setup (laptop → cluster host → kubectl), including use from Cursor Agent, see Cluster Access via SSH.