Observability

Moveat uses Grafana and Loki to inspect service logs and operational behavior.

Goals

Observability should help answer:

Is Platform healthy?
Is Agent receiving messages?
Are requests failing?
Which correlation ID belongs to a failed user action?
Are database or Redis dependencies unstable?
Are logs growing too fast?

Log flow

Correlation IDs

Platform emits correlation IDs for HTTP requests. Operations should use these IDs to connect frontend errors, Platform logs and Agent calls.

Request body logging

Platform can log request and response bodies for debugging. Sensitive fields such as passwords must be redacted. Body logging is useful during early product development, but should be controlled carefully as traffic and data sensitivity grow.

Health checks

Health endpoints should be excluded from noisy operational views where possible. They are useful for uptime but can overwhelm log panels.

Useful dashboards

Dashboard	Purpose
Moveat Observability	Service logs, errors and correlation search.
Moveat Operations	Container and operational metrics.

Common investigations

Symptom	First checks
Login returns 200 but `/me` fails	Cookie domain, CORS credentials, session storage, browser request credentials.
Agent receives webhook but no data appears	Agent Platform client logs, internal token, channel resolution, Platform write logs.
Platform is restarting	Container logs, dependency injection errors, env vars, migrations.
Grafana shows no data	Alloy status, Loki datasource, container labels, time range.