It sounds like your Prefect DB is getting bloated mostly from logs and run metadata. You can significantly reduce growth, but you cannot fully disable recording of flow/task run records — those are required for orchestration. Here’s how to minimize what’s written and how to prune what’s already there. Quick wins (Prefect 3.x) - Stop sending logs to the API - Set this on workers/runners and anywhere flows/tasks execute:

PREFECT_LOGGING_TO_API_ENABLED=false

- Or persist it to your profile:

prefect config set PREFECT_LOGGING_TO_API_ENABLED=false

- Reduce log volume - Raise the logging level (default is INFO):

PREFECT_LOGGING_LEVEL=WARNING
    # or persist:
    prefect config set PREFECT_LOGGING_LEVEL=WARNING

- Ensure results aren’t persisted - Prefect 3 defaults to not persisting results, but if you’ve changed it:

PREFECT_RESULTS_PERSIST_BY_DEFAULT=false
    prefect config set PREFECT_RESULTS_PERSIST_BY_DEFAULT=false

- Also avoid setting

persist_result=True

on tasks/flows unless you truly need it. - Avoid artifacts unless necessary - Artifacts (e.g., markdown/json artifacts) are stored in the DB and can be large. Don’t create them unless you need them. Events retention (server-side) - Events can grow quickly; set a shorter retention to keep the events table small:

PREFECT_SERVER_EVENTS_RETENTION_PERIOD="3 days"
  prefect config set PREFECT_SERVER_EVENTS_RETENTION_PERIOD="3 days"

- Docs: Settings reference What you cannot disable - Flow runs, task runs, and their states cannot be turned off; they’re required for scheduling, retries, and the UI. The best approach is retention/pruning. Prune existing data (OSS Server) - There isn’t a built-in CLI “prune” today; if you need to reclaim space immediately, you can delete old rows and VACUUM at the database level. Start with logs and events (test on a backup first!): Example PostgreSQL maintenance:

-- Delete logs older than 14 days
  DELETE FROM log WHERE timestamp < NOW() - INTERVAL '14 days';

  -- Delete events older than 7 days (match your configured retention)
  DELETE FROM event WHERE occurred < NOW() - INTERVAL '7 days';

  -- Reclaim space
  VACUUM (VERBOSE, ANALYZE) log;
  VACUUM (VERBOSE, ANALYZE) event;

- If you have very old runs you no longer need, you can delete completed flow runs (which cascades to task runs, states, and logs) via the API. If you want, I can share a small Python snippet that filters for old completed runs and deletes them in batches. A couple of clarifying questions so I can tailor this: - Are you on Prefect Cloud or self-hosted Prefect Server? - Which Prefect version (2.x vs 3.x)? I assumed 3.x above. - Do you know which tables are largest (e.g.,

log

,

event

,

artifact

,

flow_run

,

task_run

)? If you can share a quick breakdown, I can suggest a targeted plan. Helpful docs - Settings reference: Prefect 3 Settings - Database maintenance (overview): DB maintenance (guidance on pruning and VACUUM) If you confirm your setup, I’ll provide exact commands or a ready-to-run cleanup script.