Hi,
We are using pmm-server 1.17.1 in Docker. I’ve noticed in past versions that some of the associated docker volumes keep growing despite low retention settings. Usually we just remove pmm-server and pmm-data and start over, but this time I poked around a bit.
Turns out that QAN data is not being purged. We have QUERIES_RETENTION=2, yet none of the tables mentioned in /etc/cron.daily/purge-qan-data are being purged. Our query_class_metrics table was 22 GB and had records going back over 6 months (since we last did a full re-install). Also, the log file /var/log/purge-qan-data.log was empty prior to my manual run of /etc/cron.daily/purge-qan-data, so it would seem that pmm-server is never automatically executing purge-qan-data via cron.