Automated migrations with periodic metric checks

The current process of running database migrations had limitations, including relying on a single engineer to monitor metrics across multiple instances, keeping track of migration status, and only allowing migrations during office hours. To address these issues, a new approach was developed using a `metrics service` that delegates work to `metric providers`, which can be specific to one metrics source without changing the `metrics service`. The `migration runner` schedules periodic calls to the `metrics service` to determine if the migration should be stopped or not, and uses signals to control the execution of the migration. This approach allows for more efficient monitoring, 24/7 migration execution, and reduced engineer burden.