N8n update 0.210 -> 0.220 leads to memory leaks

Subj. After a week of running, i’ve connected to VDS, restarted n8n and see this. I cant open the executions page, the browser is super slow.
Is there any other stable version i should try?

My env vars:


Version 211:

  • Handle memory issues gracefully.

Well, not that gracefully :slight_smile:

But i dont see much memory issues fixes. Only leak in Node Detail view in 225.1 and 224.2.
I’ll try 226, lets see what happens

Hey @yavetal,

Have you tried the latest version of n8n? We have not seen any leaks on the backend for a long time. I am just looking at your log output as well how much memory is that saying your n8n installation has?

Not yet - and that’s actually why i’m pretty careful with updates.

Version 0.226, apparently, disabled the feature i’ve made a couple of weeks ago for my syncer: workflow activation based on the workflow name:
“Deactivating workflow during import”.

Can i disable that deactivating and keep active workflows active?

Regarding the memory leak: i’ve cleaned old n8n v0.220 instance, and now v0.226 seems to work okayish. I’ll check it in a couple of days.

Hey @yavetal,

What makes you think it is a memory leak? Is your n8n not doing anything or is it running a lot of workflows that handle a lot of data?

Browser was super slow, i was not able to open workflow executions, and there was no output (automatically created tasks in Clickup) of the workflow. Also these lines in the log:

After the reinstall and restart, everything seems to be ok. So its hard to say where exactly was the issue. Next time, probably i should save logs first.
That leads me to idea - can i set the logs output via the env vars? So they would stay even if i clean the n8n. I’ll check that.

I’ll create separate question regarding the activation in order not to mix things up

Hey @yavetal,

A slow browser would indicate an issue with the front end leaks we found and fixed recently but you don’t tend to see those in the backend logs. It looks like there are a couple of different issues going on. The front end should be a lot better in newer versions, The logs on the backend tend to mean that your n8n instance just doesn’t have enough memory but without knowing more about your setup it is hard to confirm.

Yeah, agree. Are there any ways to better report the issue next time?

The core problem was: tasks were not generated in Clickup. I went to check the n8n, connected to VDS and was not able to check the executions (slow browser).

So there’s at least one issue on the backend side i’ve faced. Should i redirect logs to be stored separately or something, to bring you more info next time?

Hey @yavetal,

Posting in here is all good :slight_smile:

You can set an option to keep the logs on disk somewhere through the env options but sometimes they are not needed like this one there are 2 parts the old leak in the front end and the instance looking like it doesn’t have enough memory for the backend.

Here we go again. Its looking like my VDS is running out of memory.
Well, it might be caused by the other users profiles, but - just saying - it started to happen after I updated from 0.210 to 0.220.

After the restart and running the workflows, memory consumption is 160 Mb. Lets see what will happen in a couple of days.

Hey @yavetal,

I would expect that error if there is not enough memory, In this case for some reason it is crashing out when trying to use a small amount of memory. This is likely to be an issue outside of n8n which is why there are not more reports of this happening.

Looking at your Windows output you are already at 85% for your memory usage so I suspect you are going to run into some issues, What is the total amount of memory your VDS has? I would be expecting at least 2GB if you are trying to run Windows as well.

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