fix: Prevent misinterpreting alembic migration errors as event loop errors

[achimnol]

This PR removes the extra exception context to check if there is no event loop running when executing alembic migrations, which makes it often confusing whether the migration failure is from the event loop handling mistake or actual migration scripts.

Before:

INFO  [alembic.runtime.migration] Running upgrade 20218a73401b -> 1d42c726d8a3, Migrate container registry config storage from `Etcd` to `PostgreSQL`
Traceback (most recent call last):
  ...
RuntimeError: no running event loop

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  ...
asyncpg.exceptions.NotNullViolationError: column "registry_id" of relation "images" contains null values

After:

INFO  [alembic.runtime.migration] Running upgrade 20218a73401b -> 1d42c726d8a3, Migrate container registry config storage from `Etcd` to `PostgreSQL`
There is no container registries data to migrate in etcd.
Traceback (most recent call last):
  ...
asyncpg.exceptions.NotNullViolationError: column "registry_id" of relation "images" contains null values

Checklist: (if applicable)

• Milestone metadata specifying the target backport version