Is this a template? Written by yourself? Language?

Besides the deployments logging is there anything else we can look at

Not a template, it's our API

It's an expressJS server

Did you make a specific change and then it started occuring?

we changed out our logger and it started happning around 7 hours ago

What is the memory spiking to? How do you know it's an OOM?

But we cannot determine anything further, everything points to us doing things more efficiently vs less efficiently

repeated Railway OOM errors and crashes

What's the limit you have set

4vCPU and 6GB memory

2 replicas

Anything that we can do to get more detailed information on Railway?

(thanks for the prompt reply btw)

Railway won't profile your application

I'm not sure how I would debug this on Javascript

yeah

Is this app open source?

no private API sorry i guess there isn't anything else we can get from railway and need to look at DD etc

Yeah it's not really Railway's responsibility here

I agree

Thanks for trying to help

Check for any added dependencies added to your project around the time the issue started happening

Run the Dockerfile locally to see if it's happening there too

If it does, https://nodejs.org/en/learn/diagnostics/memory/using-heap-snapshot

It’s a super hard problem because we think the errors were happening before but the OOM started after a huge logger refactor

Going to try out the heap snapshot thanks @astrid(irl)*

That's what you get for blindly trusing AI…

Can also git bisect