Workflow executions marked as “Crashed”, no error data saved, error workflow not triggered, auto-deactivated after upgrade to v2.1.4

Hi everyone,

I am running into a recurring issue with a production workflow where executions are marked as Crashed, but no error details are available, making it very difficult to debug.

Environment

  • n8n version: v2.1.4

  • This behavior started after upgrading from v1.x.x to v2.1.4

  • Trigger: Schedule Trigger

What happens

  • The workflow runs successfully for several loops, then suddenly stops

  • Executions appear as Crashed in the Executions panel

  • When opening the execution, I see:

    “Can’t show data. The execution was interrupted, so the data was not saved.”

  • No node shows an error

  • The configured Error Workflow is not triggered

  • After multiple crashes, n8n automatically deactivates the workflow with:

    “This workflow was automatically deactivated due to multiple crashed executions”

  • I then have to manually reactivate it

Workflow context

  • The workflow loops over items and performs:

    • External scraping

    • Data merging

    • Filtering and object creation

  • Several nodes are configured with continueRegularOutput

  • Even when disabling this during testing, the behavior persists

  • This appears to be an execution interruption rather than a normal node failure

Why this is confusing
Since the execution is interrupted:

  • No execution data is saved

  • No error message is visible

  • Error workflow does not run

  • There is no indication which node caused the crash

Questions

  1. What typically causes executions to be marked as “Crashed” with no saved data?

  2. Can this be caused by memory limits, execution time limits, or worker restarts?

  3. Is there a way to get more diagnostic information for crashed executions?

  4. Is the auto-deactivation always triggered by repeated crashes, even if they are infrastructure-level issues?

  5. Can automatic workflow deactivation be disabled or configured?
    This behavior only started after upgrading from v1.x.x to v2.1.4.

My workflows crash but i never get a crashed method.
2. This is caused by the load of your data. Add loop nodes and set batch size any number 1 through 5.

  1. On cloud, I don’t believe so on self hosted, its in the ~n8n folder somewhere in the files.
    1. I have no idea my crashed workflows never get marked or disabled.
    2. Maybe, I would look around the n8n settings interface.