Wonder if this is related to my question about logs above?

Yeah tl;dr if you're using the tool at volume although they give you suggested instance size really you need to off-board log storage to an external service. Check out how to configure logging to S3 - this then takes a lot of the disk space it's eating up on the compute instance and allows for placement/retrieval in S3. I will note this S3 logging documentation is in the Kubernetes deployment section but I believe same principals apply.

Can we simply not delete the logs periodically?

I mean yes but who wants to do that lol But real talk - why opt for a setup that increases your manual maintenance requirements and intermittent failures? For long running instances generating high log volume would expect to see external log storage and archiving/retention policy.

We don’t need logs after 30 days, so following the KISS principle here