rktechie
PROOP
4 months ago
I have a python service running on 12 vcpu and 20gb memory.
in metrics panel im using less than 1gb
please help on how to debug this. I cant find any information in the logs
3 Replies
4 months ago
It's likely there was a RAM spike. Is there any way you can attach a profiler to it? Does it happen consistently when you do one thing?
Railway's RAM polling time won't show quick giant spikes.
it errored out on 4gb and then also when i increased the memory to 20gb
it happens when i run a RAG ingestion job
4 months ago
Is there any way you can tell that job to calm down and use less RAM? I don’t know anything about what you’re doing.