Create a self-healing scheduler services (x2)

[evantahler]

Relies on Micronaut

• Clean up jobs table with TTL ([Specs] Scheduler V2 Job Retention Cron. #11869)
• Restart temporal workflows that got stuck/stopped (Automatically find failed connection manager workflows and restart them #14043)

There will likely be 2 services for each above links. Keeping separate services per job means that a failure is isolated and the other services can continue to run. There might be a "talks to our DB & services" service, and then a "talks to temporal" service.

In a future story, we will emit metrics when these jobs run.

Be sure to add an E2E test for the temporal schedule healer

[evantahler]

Note from sprint planning for @jdpgrailsdev: As the temporal work wants a scheduler, this is likely the first Micronaut story to prioritize

[davinchia]

Can we also emit a warning/metric when this happens? In theory we should never run into self-determinism errors on Temporal. It indicates a big error, and I think it's worth shouting out to us.

[evantahler]

Metrics will be emitted here #15699

[evantahler]

Once this is done, we can then merge #14758 as well

[gosusnp]

Depending on how we handle workflow versioning once we have this change ready, we may want to consider #16762