`failure_count_threshold`: We should never need this, but we don't
want to block access in tests because of this.

`failure_wait_time`: Setting it to 0 now allows each storage attempt

`storage_timeout`: Try a bit longer to access storage on CI in case
the slow machines take a bit longer to spin up the process to perfom
the check

This allows testing every storage attempt after a failure. Which could
be useful for tests

Add notes about database performance for MySQL

See merge request !13516

[ci skip]

fix confidential border issue as well as confidential styles leaking on new MR

Closes gitlab-ee#3121

See merge request !13388

Migrate force push check to Gitaly

See merge request !13406

Fixes the correct import modal window not showing

Closes #36318

See merge request !13490

Add option to disable project export on instance

See merge request !13211

Fix formatting of patch_versions.md

See merge request !13503

Better categorize test coverage results

See merge request !13485

Enable Timecop safe mode

Closes #32004

See merge request !13498

Better caching and indexing of broadcast messages

Closes #31706

See merge request !13429

Include the `is_admin` field in the `GET /users/:id` API when current user is an admin

Closes #36213

See merge request !13501

Make Git revision dropdown style consistent

See merge request !13116

Pending delete projects no longer return 500 error in Admins projects view

Closes #35435

See merge request !13389

Detect and keep track of flaky specs

See merge request !13021

Also marks a few things as uncovered, and removes an unused class.

Delete correct key from `session` after authenticating using U2F

Closes #36096

See merge request !13499

Caching of BroadcastMessage instances has been changed so a cache stays
valid as long as the default cache expiration time permits, instead of
the cache being expired after 1 minute. When modifying broadcast
messages the cache is flushed automatically.

To remove the need for performing sequence scans on the
"broadcast_messages" table we also add an index on (starts_at, ends_at,
id), permitting PostgreSQL to use an index scan to get all necessary
data.

Finally this commit adds a few NOT NULL constraints to the table to
match the Rails validations.

Fixes gitlab-org/gitlab-ce#31706

Use a separate table for storing push events

See merge request !12463

Cache Appearance instances in Redis

Closes #36066 and #31698

See merge request !13433

Show error message for API 500 error in tests

Closes #36174

See merge request !13502

Do not run the `ee_compat_check` job for stableish branches

Closes #35131

See merge request !13497

Signed-off-by: Rémy Coutable <remy@rymai.me>

Signed-off-by: Rémy Coutable <remy@rymai.me>

Signed-off-by: Rémy Coutable <remy@rymai.me>

Signed-off-by: Rémy Coutable <remy@rymai.me>

Signed-off-by: Rémy Coutable <remy@rymai.me>

[skip ci]

Signed-off-by: Rémy Coutable <remy@rymai.me>

document have_gitlab_http_status

Merge branch '36328-deserializemergerequestdiffsandcommits-fails-due-to-null-too_large-column-9-5-rc1' into 'master'

Fix merge request diff deserialisation when too_large was absent

Closes #36328

See merge request !13500

Enable the RSpec/HookArgument cop and auto-correct offenses

See merge request !13484

Remove an unnecessary `let` in spec/features/projects/user_edits_files_spec.rb

Closes #36308

See merge request !13488

Bumps omniauth-ldap gem version to 2.0.4

Closes #25637

See merge request !13465

Not all diffs in st_diffs had the too_large attribute set at all, but the column
is non-nullable. Explicitly cast to boolean to avoid that problem.