From time to time we find our API service will start hanging. It will accept requests but then never returns. We suspect connections to postgres or redis are hanging but we're not certain. Restarting fixes the issue but it's unclear why it happens in the first place.

We can ssh into the instance but without tools like lsof/ss/strace available it's hard to understand what's going on.

Does the team have any advice for debugging such issues on railway?

Worth pointing out our production environment has been moved back to google whilst the disk issue with metal is being fixed.