logrotate best practice

Hi, I recently had a critical issue at a customer where the nodes kept shutting down due to full catalog volume. Somewhere along the way - probably to solve another issue - they had turned log level to DEBUG for some components and were generating ~2 GB of gzip logs daily - not an issue, until they had accumulated 90 days worth! We noticed then that the default logrotate keeps logs for a very long time - looks like up to 52 weeks. Did any of this change recently - default log lovel, or logrotate config? I've never seen this type of log history onsite, so are customers changing this, cleaning manually, or I just never noticed because scrutinize doesn't usually pack up old logs unless configured to do so? (Support fixed the issue by resetting log levels to INFO and deleting old logs.)