Connection lost when running 'Edit Image' node

Questions
1

Describe the problem/error/question

I am building a new workflow that has Edit Image node. When it reaches the edit image node (resize) the ‘Connection lost’ message appears in top right corner. I restarted my PC as well as updated n8n to latest version + restarted again the n8n workspace from the admin panel. I’m running the cloud version. I have other workflow with image nodes but they work fine. How can I identify which of the nodes might create some corruption/ issues in the workflow? I also copy pasted all the nodes in a new workflow but get the same error. If I let it run it eventually resets the entire workflow and starts from scratch. I am converting the image to webp before the image node, so it’s probably not an image size issue. Might be a corrupt node somewhere in the workflow, not sure.

What is the error message (if any)?

n8n may have run out of memory while running this execution. More context and tips on how to avoid this in the docs

Please share your workflow

Share the output returned by the last node

Information on your n8n setup

  • n8n version: 2.18.5 (Cloud)
  • Database (default: SQLite):
  • n8n EXECUTIONS_PROCESS setting (default: own, main):
  • Running n8n via (Docker, npm, n8n cloud, desktop app):
  • Operating system: Windows 10
3

Hi @andrei.turca as the error says, n8n cloud space has ran out of memory, i think if you are processing good quality or huge size images, then consider getting yourself a higher tier plan, or you can actually self host if you want but, currently upgrading your plan is the only option, as self hosted instances have some config options like saving big files into file storage instead of memory, but in cloud there is nothing like that, so consider upgrading your plan.

→ N8N Cloud Vs Self-Hosted » Qixazow

4

Thanks for the quick reply.

I was only trying to edit 1 webp image.

I have other workflows with many edit image nodes that receive raw images. I tested those and they run without issues.

Update:
I changed the webp image resolution to a lower one before sending to the edit image node. It seems that solved the issue.

Probably the high resolution was behind the error. I thought the image disk size might cause issues. That’s why I converted to webp before Edit Image node. But it seems is the resolution not the image disk size.

5

Hi @andrei.turca, do the resize on wsrv.nl itself by adding w and h query params — it returns the already-resized webp so the Edit Image node never has to load the full-res buffer into n8n’s memory, which is what’s blowing the cloud worker.

swap your existing Convert image node for this and drop Edit Image entirely, saves a node and the OOM.