This topic was automatically generated from Slack. You can find the original thread here.
Is there a good way to debug memory issues with Pipedream webhooks? I have a workflow which is consistently having an out of memory alert.
Not yet sorry, that’s something we plan on improving!
Ok, is the only option to increase memory? Seems to happen when there is a burst of traffic to Pipedream.
Yes, try increasing a little bit each time and see if it stops failing
If it’s linked to high traffic, it could have to do with reused instances & uncollected garbage adding up (effectively a memory leak).
Yeah, it exclusively happens when we have high traffic.
This would be a typical run, only one failure but the workflow is somewhat business critical for the client.
How long are the other runs taking?
We have a filter so most of them will be filtered and take less than 500ms. Those that are running fully take between 20 and 30s.
I see, thanks. Is your timeout set to 30s?
90 seconds, we set that when we got a few timeouts, but now we have the memory issue.
Ah right, if it’s set at 256MB try 512MB, it might help
Thank you, I’ve increased it. Only issue is that it doubles our credit usage and the flow overall is quite simple, so cost becomes a factor but we’ll monitor and see.
Totally get you. Good thing credit overage is very cheap. If your client is on the Advanced Plan you can set up auto-retries on errors, then you can keep it at min memory