Automatically update MR merge-ref along merge status

See merge request gitlab-org/gitlab-ce!29569

Disallow `NULL` values for `geo_nodes.primary` column

Closes gitlab-ee#12061

See merge request gitlab-org/gitlab-ce!29818

Migrate managed clusters that aren't using managed features to unmanaged

Closes #62772

See merge request gitlab-org/gitlab-ce!29251

There are clusters that have Kubernetes namespaces
stored which are missing a service account token.
These namespaces are unable to be used for deployments,
so marking the clusters as unmanaged will allow the
platform credentials to be used instead.

This couples the code that transitions the `MergeRequest#merge_status`
and refs/merge-requests/:iid/merge ref update.

In general, instead of directly telling `MergeToRefService` to update
the merge ref, we should rely on `MergeabilityCheckService` to keep
both the merge status and merge ref synced. Now, if the merge_status is
`can_be_merged` it means the merge-ref is also updated to the latest.

We've also updated the logic to be more systematic and less user-based.

In https://gitlab.com/gitlab-org/gitlab-ce/merge_requests/21497, we
migrated all project import data into a separate table,
`project_import_data`.  In addition, we also added:

```
ignore_column :import_status, :import_jid, :import_error
```

In https://gitlab.com/gitlab-com/gl-infra/production/issues/908, we
observed some of these `import_error` columns consumed megabytes of
error backtraces and caused slow loading of projects whenever a `SELECT
* from projects` query loaded the row into memory.

Since we have long migrated away from these columns, we can now drop
these columns entirely.

Add migrations needed to encrypt feature flags client tokens

See merge request gitlab-org/gitlab-ce!29320

Add index to events and audit_events tables

Closes gitlab-com/Product#309

See merge request gitlab-org/gitlab-ce!29194