This is a temporary fix for the in-product-marketing-email worker.
It is a bug that consists of multiple layers. We expect to only have
root namespaces in `onboarding_progresses`. However, groups can be moved
to sub-groups. Threfore we also have sub-groups that would receieve the
emails.

This alone would not be an issue. But we need to call `.root_ancestor`
for these sub-groups and it doesn't play well together with
`each_batches` when there is an other JOIN condition on the namespace.
It currently generates a non-working query that fails.

To prevent this for now we do not send emails to sub-groups as a
temporary solution. In the long run, we need to remove all sub-groups
fron `onboarding_progresses` and fix the issue of calling
`.root_ancestor` in an `each_batches` call.

Use gitlab-fog-google instead of upstream fog-google gem

See merge request gitlab-org/gitlab!56245

Fix mismatched custom vue event names

See merge request gitlab-org/gitlab!55958

Ensure that the custom emitted events for
the type of work chart labels filters are
consistent

Create single component for Pipeline Mini Graph

See merge request gitlab-org/gitlab!55199

Fix RuboCop offenses for Style/ClassEqualityComparison

See merge request gitlab-org/gitlab!56227

Clean up anchor in Backup docs

See merge request gitlab-org/gitlab!56363

Update two autodevops template references to default branch

See merge request gitlab-org/gitlab!56304

Merge branch '300351-es-migration-add-permission-data-to-notes-attempt-3-as-indexed-json-changes' into 'master'

Delete orphaned notes from Elasticsearch index

See merge request gitlab-org/gitlab!56162

Move recursive namespace query to INNER JOIN [RUN ALL RSPEC] [RUN AS-IF-FOSS]

See merge request gitlab-org/gitlab!56078

Add comments to Jira issues

See merge request gitlab-org/gitlab!56069

Fix typo in Workhorse readme

See merge request gitlab-org/gitlab!56366

Update docs for retarget MRs

See merge request gitlab-org/gitlab!56335

This ensures that:
- it is ready for production use
- documents `remove source branch` caveat

Move expectation out of `before` block

See merge request gitlab-org/gitlab!56268

Update to re-add QA selectors

See merge request gitlab-org/gitlab!56324

Ensure Project Approvals attributes (and API) casts to boolean

See merge request gitlab-org/gitlab!55492

Fix autoloading issue for SidekiqMiddleware::SizeLimiter::Client

See merge request gitlab-org/gitlab!56120

Fix tier badging for Auth0 integration

See merge request gitlab-org/gitlab!56364

Add allow force push docs

See merge request gitlab-org/gitlab!55751

Added docs for the allow force push feature

Add Allow force push option to Protected branches

See merge request gitlab-org/gitlab!55261

Get vulnerabilities from db [RUN ALL RSPEC] [RUN AS-IF-FOSS]

See merge request gitlab-org/gitlab!55641

Issue https://gitlab.com/gitlab-org/gitlab/-/issues/321086
We get vulnerabilities from database. Before that
we have to parse report on each request

Move Workhorse development to main GitLab repository

See merge request gitlab-org/gitlab!54941

Fix pipeline status when DAG jobs needs manual jobs [RUN ALL RSPEC] [RUN AS-IF-FOSS]

See merge request gitlab-org/gitlab!53476

We achieve this with skipping DAG jobs when it needs a manual job.

Since we now raise an error if noteable is `nil`
when `as_indexed_json` is called, the noteable
object will always be available when adding
project feature access level.

This relates to some more possible edge cases we may have in our data.
Since we are running a migration to backfill all these settings that
will run forever unless a value is set it is better to always have some
fallback that isn't `nil` as the migration will never finish.

In this case we can be sure that anything that is required to have a
project but doesn't for whatever reason shouldn't show up in searches
and therefore it's safe to just set `*_access_level = DISABLED`.

Prior to this we were specifically targetting missing commits because
this was a case we [observed in
staging](https://gitlab.com/gitlab-org/gitlab/-/merge_requests/55878)
but we also may have more cases we haven't run into yet for other
noteable types. The `noteable_id` column on the table does not have a
`NOT NULL` constraint nor does it have foreign key references so it's
totally conceivable there are other kinds of orphaned notes (eg. issues
that were hard deleted at some point) that we don't know about. I have
concerns that any one of these situations could again lead to the
migration to update permissions getting stuck forever. As such I think
the best course of action is to delete any of these. I also can't think
of any good reason to keep the orphaned notes in the index so this seems
quite safe.

We discovered at
https://gitlab.com/gitlab-org/gitlab/-/merge_requests/55878 that there
are sometimes orphaned commit notes. Since these are making it difficult
to migrate as we can't set the permissions correctly for these and also
since you can't actually navigate to a note on a deleted commit there is
no point in keeping these in the index.

This commit was extracted from a combined MR at
https://gitlab.com/gitlab-org/gitlab/-/merge_requests/55883 so it could
be merged independently from the migration.

Remove latest_builds_report_results preloading in pipeline serializer

See merge request gitlab-org/gitlab!56181

Update tier badge to indicate self-managed.

Clarify the use of registry_api_url

See merge request gitlab-org/gitlab!56106

Apply new GitLab UI for buttons on GitHub Import page

See merge request gitlab-org/gitlab!56096

Fix allow settings for maven duplicates

See merge request gitlab-org/gitlab!56100

Add info column into security_scans table

See merge request gitlab-org/gitlab!55983