Skip to main content

​
Before you start

  • Open a terminal in the directory containing your **docker-compose.yml**.
  • Ensure Docker Desktop / Docker Engine is running.
  • If you’re unsure of service names, run:
docker compose config --services
or to see what’s running: 
docker compose ps

​
Service names you can target

Use any of the following service names in place of api in the examples below: 
db
rabbitmq
redis
celery_beat
celery_worker
converter
web_init
web
compare
server

​
Stream live logs (follow mode)

Stream logs for a specific service: 
docker compose logs -f web
Press Ctrl+C to stop following. Common filters 
# Only the last 200 lines
docker compose logs --tail=200 web

# Since a relative time window (e.g., last 2 hours)
docker compose logs --since=2h web

# Since/until absolute timestamps (ISO 8601)
docker compose logs --since="2025-09-30T01:00:00" --until="2025-09-30T03:00:00" web

# Include timestamps in each line
docker compose logs -f --timestamps web
All services at once 
docker compose logs -f

​
Export logs to files (Linux/macOS)

Entire history (can be large): 
docker compose logs web > web.log
Recent logs with timestamps (recommended for support): 
docker compose logs --since=24h --timestamps web > "web-$(date +%Y%m%d-%H%M).log"
Compress for sharing: 
gzip -9 web-*.log
# produces e.g. web-20250930-1042.log.gz
Stream & save simultaneously: 
docker compose logs -f --since=30m --timestamps web | tee "web-live-$(date +%Y%m%d-%H%M).log"

​
Export logs for all services (copy/paste)

​
Linux/macOS

services=(db rabbitmq redis celery_beat celery_worker converter web_init web compare server)
stamp=$(date +%Y%m%d-%H%M)

for s in "${services[@]}"; do
  docker compose logs --since=24h --timestamps "$s" > "${s}-${stamp}.log"
done

tar -czf "api-selfhosted-logs-${stamp}.tar.gz" *-"${stamp}".log

​
Per-container logs (scaled services)

If a service is scaled (e.g., web has multiple replicas), you can export each container’s logs: 
# List running containers and their service
docker compose ps

# Export per-container (replace with actual container name, e.g., stack_web_1)
docker logs --since=24h --timestamps stack_web_1 > stack_web_1.log
Container names are typically <project>_<service>_<index>. For ad-hoc viewing, docker compose logs <service> is simpler; for per-instance analysis, use docker logs <container>.

​
What should I collect for Support?

When requested, please provide:
  1. The .log, .tar.gz, or .zip files you created above (ideally limited with --since and/or --tail).
  2. The time window during which the issue occurred.
  3. The service name(s) involved (e.g., web, compare, server).
  4. Any noteworthy steps just before the issue (deploy, restart, config change, etc.).

​
Troubleshooting tips

  • Keep it concise: Use --since (e.g., 2h, 24h) and --tail to reduce noise and file size.
  • Timestamps help: Add --timestamps so lines can be correlated across services.
  • Startup issues: Check web_init and server first; they often surface misconfigurations early.
  • Dependencies: If the app looks healthy but jobs stall, review celery_worker and celery_beat. For connectivity or persistence problems, check db, redis, and rabbitmq.
  • Redaction: Logs may contain URLs or IDs. Skim before sharing externally.

​
FAQ

Yes. docker compose resolves services from the docker-compose.yml in the current directory (or use --file to point to the correct one).
docker compose logs prints plain text from the Docker log driver. If your services emit JSON lines, they’ll appear as such. For structured querying, consider a central log stack (e.g., Loki/ELK) in addition to these ad-hoc exports.
Narrow with --since, --until, and --tail. For example:
docker compose logs --since=90m --tail=1000 --timestamps web > web-window.log
Get in touch with our support team at support@draftable.com and we will do our best to help you.