On a have node, with a lot of users, sometimes mysql connections drop, causing rendering errors

Sometimes mysql connection drop, where the client tells us mysql is gone, an the server tells an error message. This causes unexpected behaviour for the end-user during login etc:

kubectl logs -n stackspin single-sign-on-database-mariadb-0

2023-08-30  1:59:54 396767 [Warning] Aborted connection 396767 to db: 'stackspin' user: 'stackspin' host: '10.42.0.88' (Got timeout reading communication packets)
2023-08-30  2:01:50 379803 [Warning] Aborted connection 379803 to db: 'kratos' user: 'kratos' host: '10.42.0.35' (Got timeout reading communication packets)
2023-08-30  2:01:50 396792 [Warning] Aborted connection 396792 to db: 'stackspin' user: 'stackspin' host: '10.42.0.88' (Got timeout reading communication packets)
2023-08-30 11:56:46 409704 [Warning] Aborted connection 409704 to db: 'stackspin' user: 'stackspin' host: '10.42.0.88' (Got an error reading communication packets)
2023-08-30 11:58:28 409727 [Warning] Aborted connection 409727 to db: 'stackspin' user: 'stackspin' host: '10.42.0.88' (Got an error reading communication packets)
2023-08-30 12:07:21 408153 [Warning] Aborted connection 408153 to db: 'stackspin' user: 'stackspin' host: '10.42.0.88' (Got an error reading communication packets)
2023-08-30 12:07:21 408763 [Warning] Aborted connection 408763 to db: 'stackspin' user: 'stackspin' host: '10.42.0.88' (Got an error reading communication packets)

It is mostly noticed directly after restarting the dashboard.