Production Checklist

Use this checklist before production deployment.

Build and Configuration

Run forj build.
Confirm .goforj.yml matches intended components.
Confirm named apps listed under apps have the intended component choices.
Confirm *_SUPPORTED_DRIVERS includes all runtime drivers.
Confirm secrets are provided by the deployment environment.
Confirm APP_ENV and APP_DEBUG are production-safe.

Runtime Topology

Choose standalone or split runtime processes.
Run HTTP, workers, and scheduler through explicit commands.
For named apps, run the app-specific binary such as ./bin/marketplace api.
Ensure scheduler singleton or locking behavior is correct.
Set shutdown timeouts deliberately.

Data

Run migrations intentionally.
In multi-app Projects, confirm which app owns each migration stream.
Verify database readiness.
Confirm cache is not source-of-truth storage.
Confirm storage disks and object storage permissions.
Verify forj build has refreshed generated named resources.
Run forj backup:plan and confirm every discovered database and storage resource has the intended strategy.
Verify a recent backup and test its restore path in an isolated environment.

Observability

Verify health and readiness endpoints.
Configure APP_DIAG_TOKEN.
Verify metrics scrape targets.
Confirm metrics, logs, inspects, and Lighthouse preserve app identity.
Confirm high-cardinality labels are not introduced.
Confirm logs are high-signal and secrets are not logged.
Enable Lighthouse only where appropriate.

Async

Confirm job handlers are registered.
Confirm workers use the intended queue driver.
Confirm retry and idempotency behavior.
Confirm queue shutdown timeout.
Confirm event drivers match process topology.

Final Check

Run smoke tests.
Inspect route:list.
Validate one health check.
Validate one readiness check.
Validate metrics output.
Validate expected runtime processes start and stop cleanly.
Confirm backup creation, verification, retention, and freshness monitoring are automated.

Next Steps