Do this TODO message for SaaS and on-prem.

We are creating a DeletedObject for every Metric and Cinstance that is destroyed in System indicating its Service, and when the Service is destroyed, one of the actions it does is fetching all these metrics and cinstances of that Service (because we have the service_id) to destroy the Stats in Backend.

The problem now is that we never destroy entries of DeletedObject right now even when we won't need them anymore after they are used and this will end up growing too much.

The goal of this Jira issue is to regularly destroy (through the Janitor) all those DeletedObject whose owner (the service) has been destroyed a while ago, maybe 1 week after the Service has been destroyed seems safe enough.