Why does n8n display stale 'running' executions even after workflows have completed?

It sounds like your executions list is showing stale data—executions that are marked as running even though they’ve already finished. This can happen for a couple of reasons:

1. Execution Data Saving Settings:
   n8n allows you to tailor which execution data gets saved (and when). If you have settings that save data on progress or for manual executions, you might end up with extra records if the execution state isn’t updated as expected. Review your environment variables (for example, EXECUTIONS_DATA_SAVE_ON_PROGRESS or EXECUTIONS_DATA_SAVE_MANUAL_EXECUTIONS) to see if they align with your expected behavior.

2. Data Pruning Configuration:
   When executions finish, they aren’t deleted immediately. Instead, n8n holds onto them for a set period (default is 14 days) or until the maximum count is reached. If automatic pruning isn’t configured (using EXECUTIONS_DATA_PRUNE, EXECUTIONS_DATA_MAX_AGE, and EXECUTIONS_DATA_PRUNE_MAX_COUNT), old executions might continue to appear even if the workflow itself finishes quickly.

3. Database Specific Behavior (SQLite):
   When using SQLite as your database, even if executions are pruned, the freed space isn’t automatically returned to the operating system unless you run a VACUUM (or set DB_SQLITE_VACUUM_ON_STARTUP). While this is more about disk space than the “running” status, it could hint at how data isn’t being fully refreshed.

Regarding the EXECUTIONS_PROCESS setting, its value depends on your Docker configuration. Since you’re not sure what is being used, check your Docker run command or Docker Compose file for any environment variable settings related to execution processing. If there’s no explicit setting, n8n defaults will be used.

Next Steps

• Review your Environment Variables:
  Check the settings related to the execution data—for example, EXECUTIONS_DATA_SAVE_ON_ERROR, EXECUTIONS_DATA_SAVE_ON_PROGRESS, etc.—and adjust them if you don’t need to persist intermediate states.

• Enable and Configure Data Pruning:
  Consider enabling automatic data pruning or reducing the retention period so that finished executions are removed and don’t clutter the view. For instance:

  SHCopied
  # Activate automatic data pruning with a shorter max age (in hours)
  export EXECUTIONS_DATA_PRUNE=true
  export EXECUTIONS_DATA_MAX_AGE=24  # Adjust based on your needs

• Verify Your Docker Setup:
  Look into your Docker run command or Docker Compose file to determine whether the EXECUTIONS_PROCESS setting has been overridden.

By fine-tuning these settings, you should be able to reduce the occurrence of stale “running” executions appearing in your dashboard. If problems persist, checking the server logs might help diagnose if there are any issues updating the execution state.