Skip to content

Monitoring & troubleshooting

  • Liveness: GET https://<platform-host>/api/health{"status":"ok"}.
  • Post-change verification: ./sprig azure smoke re-checks resources, Entra wiring and the health endpoint.

The platform Function App writes to Application Insights (requests, failures, dependency calls) backed by a Log Analytics workspace — standard KQL applies. The audit trail (business events: publishes, shares, decisions) lives in the platform storage tables and is queryable independently of telemetry.

Symptom Likely cause Fix
Builder gets not onboarded No metadata row / sandbox ./sprig azure onboard <upn>
Publish fails at app registration Entra propagation lag Retry; the flow is idempotent
App deploys but sign-in loops Auth settings missing on the Static Web App Re-publish (first publish sets them); check SPRIG_AAD_* app settings exist
Health endpoint cold and slow Consumption cold start Expected on first hit (~30 s)
Share approval succeeds but viewers still blocked Entra group assignment propagation Allow minutes; verify the group on the app’s service principal
Viewer opens app, sees empty data Viewer lacks Fabric permissions Working as designed — grant data access via Fabric, not sprig

Each published app’s registration and resource carry sprig-owner / sprig-app tags — start from the resource, identify the owner, pull the audit trail for that app from the events table. The trail names every actor and operation with timestamps.