This topic was automatically generated from Slack. You can find the original thread here.
I have a workflow that receives a call to an endpoint that is giving Out Of Memory errors at trigger consistently. I have upped it to 512MB and that hasn’t helped. I don’t understand what could be using so much memory. Is there anyway to diagnose? The payload is not that large, and contains no binary data.
If I go in and replay the event, it works fine every time.
The workflow has 9 steps. All of which are either text manipulations or BigQuery SQL calls.
I’m having the same problem. It looks like its getting stuck at the trigger, which mine is just to check if its 9AM, and grab the time.
even weirder is mine ran a second time on its own and finished successfully. 
I have my Memory set to 512 and the largest thing I work with in this is only 20MB everything else is just text from a google sheet. So I should never run out of memory
Now another workflow of mine which has worked consistently for weeks has started throwing Out Of Memory errors.
I feel like something is wrong for sure.
@U06JJAV16GL could you both please share your projects with support and then share here the workflow’s URL so I can take a look
I’ll look at the workflows and let you guys know! Not sure when I’ll have any additional info
I have disabled it for now, as nothing was working.
So there is no impact, let me know if I can do anything to help!
for your workflow, I see that you’re downloading a video to /tmp. Since that loads the whole video into memory, that should be related to the OOM error
Even though it seems like the workflow didn’t execute at all, it definitely did start running and broke in the middle of the execution
Could you try raising the memory even more or try with a smaller size video?
The video it tried to download is 36.8MB, and I have memory set to 512MB
The total memory used by the execution environment includes memory usage from many different parts
I wonder if you increase even more the memory if it executes successfully
It should run at 2PM, so I’ll increase it to 768 and see what happens.
