[MULTIPLE BUGS?] Workflow infinite "running"

Describe the problem/error/question

I’m noticing one particular workflow triggered by a GMAIL trigger (when a new email comes in every minute). I’m seeing infinite execution running forever…


My concern is I have seen a scenario where an “email” got stuck until I manually deleted that execution and then a new trigger started and “got” that email that was stuck. I don’t know why this is happening but the most important issue is that I have a max 5-minute timeout:

Information on your n8n setup

  • n8n version: 0.221.2
  • Database (default: SQLite): SQLite
  • n8n EXECUTIONS_PROCESS setting (default: own, main): own
  • Running n8n via (Docker, npm, n8n cloud, desktop app): docker/caddy
  • Operating system:

Today again I ran the same issue where an email was stuck and wasn’t processed properly.

Now I’m using version: 0.222.3

Also, I’m starting to see this [UNSAVED WORKFLOW] thing in my execution list:

I am also on version 0.222.3

When debugging some stuff today, I also noticed those [UNSAVED WORKFLOW] workflows running, even though afaik I did not have any unsaved workflows open.

Definitely, something very buggy is happening.

More stuff:

I’m trying to “execute” a test webhook:

Where I Start the trigger event and then send the request to the test URL. I know it works, because I got the email into my gmail run by the process, but the window is never showing up what happened in that event.

The stop button is not working either.

So the only way to stop it is by leaving the current workflow window.


just a screenshot as i saw it happening right now, Version 0.222.3

Hi folks, I am sorry for the trouble.

I feel like there are multiple things mixed into a single thread here.

I’m noticing one particular workflow triggered by a GMAIL trigger (when a new email comes in every minute). I’m seeing infinite execution running forever…

We updated our execution data a lot recently. Can you upgrade to the most recent version of n8n 0.225.1 and confirm whether the issue still persists for new executions started after the upgrade? If so, could you confirm how exactly it can be reproduced from scratch?

Where I Start the trigger event and then send the request to the test URL. I know it works, because I got the email into my gmail run by the process, but the window is never showing up what happened in that event.

This sounds like your n8n instance can’t pass on execution data to your browser. Such problems typically come down to your reverse proxy configuration (or any other component sitting between your n8n server and browser). You’d need to make sure it passes on server-sent events as it should. If this typically works for a while after opening a new browser, but then suddenly stops you would want to make sure your server keeps sessions alive for as long as needed.

I will try tonight.

This started to happen recently after upgrading to latest version. 0.222.3
I have been running this particular server for almost 6 months with daily tests and scenario modifications so I can quickly detect when something is not working as “always” or different. It’s not a new configured server so I’m not sure why this problems would start happening with no changes from my side.

@MutedJam and what about the [UNSAVED WORKFLOW] thing? Me and @FelixL have seen this in our executions.

what about the [UNSAVED WORKFLOW] thing

I am not sure I understand the problem with this one.

This seems to be expected whenever you (or anyone else on your instance) is running a workflow which has not been saved yet (or which simply has no name), for example while building a new workflow or when executing a sub-workflow from a file.

I did not have any Unsaved workflows running when I saw this.

Maybe it showed a workflow which was running a few minutes ago. :man_shrugging: