Deleting Executions in n8n Cloud

Questions
1

Describe the problem/error/question

I’m experiencing degraded performance and I think it’s when my execution logs start to get to or above the 25k limit for my plan. I can’t start new executions, delete old ones, save changes to workflows when this happens. Basically I’ve ground to a halt.

I think I can partly address this for the future by turning off executions for some of my sub-workflows, but there are a few that I’d rather be able to retain for debugging and just delete them sooner. Also I need a way to delete now so I can get my workspace operating again.

  • Because I’m on the cloud version I don’t think I can access the SQLlite database directly to delete executions like is suggested on the forum in posts like this (What is the best method to delete the executions?)
  • when it comes to performing bulk deletions from the interface I’m running into the same problem as this post ( Cannot delete executions with filters) and so can only delete 10 at a time
  • I’ve looked into using the n8n node or API but these seem to only allow execution by id, so I’d first need to get all the executions, filter out the old and then delete. The get requests seem to always time out if you’re requesting more than a handful of executions.

Any suggestions for a cloud user? Is there a way of accessing the SQLite database that I’m not aware of?

I feel the most elegant solution would be improving the n8n node/API to allow targeted deletions (e.g. based on workflow id and date range) as then any user could customise their execution log maintenance within a workflow. But maybe there are limitations with SQLite that make this impossible. Also as impressive as the steady progress on updates to n8n is, I can’t afford to wait for that to be developed.

What is the error message (if any)?

This is the error messages I get when I try
image

Information on your n8n setup

  • **n8n version: 1.6.4.3
  • **Database (default: SQLite): SQLite
  • **n8n EXECUTIONS_PROCESS setting (default: own, main): main (starting to customise)
  • **Running n8n via (Docker, npm, n8n cloud, desktop app): cloud
  • **Operating system: Windows
2

Seems I was mistaken about the n8n node. When you try to filter the get executions to a specific workflow it does time out … but if you retrieve all executions all workflows it doesn’t! It took around 3 minutes to retrieve about 15,000 executions, which means you can then create a logic inside a workflow for which ones you want it to delete before sending to another n8n node for delete by id.

I managed to get it to delete around 5,000 executions in one batch though not without error at the end. In the attempt it appears I came up against rate limits, though couldn’t find anywhere what they are.

Anyway just sharing since other cloud users have reported this issue and this might work for them too.

3

Hey @jamescv,

Another option is to go to the executions page and you can select all in there and a delete option will appear.

4

okay I take this back. I’m getting error code 524 when I try to retrieve the executions now.

5

Thanks Jon, yes that’s there as an extreme. But if I don’t want to delete the most recent ones it’s not quite ideal

6

This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.