Add the etcd defrag job to our clusters

[primeroz]

Motivation

Follow up to #1659

Vintage has a etcd3 degrag systemd service and timer https://github.com/giantswarm/k8scloudconfig/blob/master/pkg/template/master_template.go#L250

• Should we add this to capz ? this other issue might be relevant Implement ETCD auto-compaction #2089
• If we do, would it be better to just use a cronjob APP and deploy it through default-apps-azure ?

Other Info

• There is a slack conversation about running defrag more often because it does make a huge difference https://gigantic.slack.com/archives/C02HLSDH3DZ/p1678400279018599
• relevant issue in vintage https://github.com/giantswarm/giantswarm/issues/26218

TODO

• Implement the jobs
• make sure it's running properly (test :))

Outcome

For Clippy

• We have a more reliable etcd3 cluster