Operations runbook

Status and logs

docker compose ps
docker compose logs --tail=200 api kong hapi-fhir-jpaserver-start
docker compose logs --since=30m fhir-transformer hapi-fhir-postgres kong-postgres

Add --follow only during an attended investigation; logs may contain identifiers and must be handled as sensitive operational data.

Restart one service

docker compose restart <SERVICE>
docker compose ps <SERVICE>
docker compose logs --tail=100 <SERVICE>

Verify the service endpoint after restart. Restarting a database can interrupt dependent services.

Stop and start

docker compose stop
docker compose start

stop preserves containers and data. Do not use docker compose down -v; -v can remove managed volumes. This stack also uses host bind-mounted data directories, which require separate backup handling.

Reinitialize Kong

First query existing services and routes. The shell scripts use POST and are not idempotent, so rerunning them may produce 409 conflicts. Run them only to create missing objects, then repeat verification.

Incident handoff

Capture time/timezone, source revision, image digests, docker compose ps, sanitized logs, last known healthy time, recent configuration changes, and failed verification step. Never paste secrets or patient payloads into tickets.