6. Start and inspect the core services

Keep the workspace and reset profiles disabled for the fresh core deployment. Check the resolved services before startup.

sudo docker compose --env-file deploy/postgres/.env \
  -f deploy/postgres/compose.yaml --profile v2 config --services

sudo docker compose --env-file deploy/postgres/.env \
  -f deploy/postgres/compose.yaml --profile v2 up -d --wait

sudo docker compose --env-file deploy/postgres/.env \
  -f deploy/postgres/compose.yaml --profile v2 ps

The core list contains backup-volume-init, db, platform, router, web and worker. The backup directory initializer is a one-shot service and should finish successfully. Verify the long-running services report healthy where they have a health check, then run the access boundary checks.

If a listener stays unhealthy, check its database role and the identity provider's HTTPS signing-key endpoint. The readiness probe checks both. An unsigned request can return HTTP 401 while readiness still fails.

Once an authorized operator has created the intended workspace, replace the placeholder in EPM_PG_WORKSPACE_ID with its exact ID. To run its optional scoped listener, revalidate the configuration and run up -d --wait workspace using --profile v2 --profile workspace and the same environment and Compose file. The listener binds host loopback port 5192.

6. Start and inspect the core services

Discussion