Tips & tricks

Monitoring

You could implement alerts on malfunctioning Database resources by watching Database events. You can find a complete list of Reasons and Messages here. Alternatively, if you find this too granular, you can simply watch the .status.conditions[*].status.type: Ready field and check whether .status.conditions[*].status.status equals "False" for a certain number of time, if it does, that could generate an alert.

Credential rotation

Credential rotation could be performed periodically by using a simple CronJob for each Database resource in the cluster, see the bitnami/kubectl Docker image.

Restoring resources

If something bad happened and you've lost all your Database resources, you can simply reapply all your Database yaml files. Given that the create and rotate operations were implemented according to the specification, Database resources will be regenerated without the need of manual intervention.