Recently, we have seen issues with Quarry and labsdb1011 (T247978 T246970)
Labsdb1011 is always very overloaded (https://grafana.wikimedia.org/d/000000377/host-overview?orgId=1&refresh=5m&var-server=labsdb1011&var-datasource=eqiad%20prometheus%2Fops&var-cluster=mysql&from=now-24h&to=now&fullscreen&panelId=3) and we've seen in s8 that 10.4 and Buster seems to be performing better, CPU-wise.
Let's try to reimage this host with Buster and see what happens.
/srv won't be formatted.
The rollback plan would be to reimage to stretch and reclone from labsdb1012.
Once the host has no traffic, let's do a disk performance benchmarking to make sure its IOPS are similar to labsdb1012 ones.