We need this because `expect_any_instance_of` doesn't work on
prepended models. Now we could use the same code between CE/EE

Adding the :prometheus tag to an rspec test will clear out
memory-mapped files and reset the registry.

Closes https://gitlab.com/gitlab-org/gitlab-ce/issues/39968

Direct disk access is done through Gitaly now, so the legacy path was
deprecated. This path was used in Gitlab::Shell however. This required
the refactoring in this commit.

Added is the removal of direct path access on the project model, as that
lookup wasn't needed anymore is most cases.

Closes https://gitlab.com/gitlab-org/gitaly/issues/1111

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

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

Closes gitaly#1110

FactoryBot's build_stubbed ignores the current database sequence of the
projects table and starts at ID 1000. If more than 1000 projects are created
during a test run, leftover repositories can cause spec failures. For example,
a spec that expects an empty repository may fail since there may be existing
content.

Closes #5461

Closes #28857

E.g. https://gitlab.com/gitlab-org/gitlab-ce/-/jobs/52873034

I've followed the [upgrade guide](https://github.com/thoughtbot/factory_bot/blob/4-9-0-stable/UPGRADE_FROM_FACTORY_GIRL.md) and ran these two commands:

```
grep -e FactoryGirl **/*.rake **/*.rb -s -l | xargs sed -i "" "s|FactoryGirl|FactoryBot|"
grep -e factory_girl **/*.rake **/*.rb -s -l | xargs sed -i "" "s|factory_girl|factory_bot|"
```
Signed-off-by: NRémy Coutable <remy@rymai.me>

Moving the check out of the general requests, makes sure we don't have
any slowdown in the regular requests.

To keep the process performing this checks small, the check is still
performed inside a unicorn. But that is called from a process running
on the same server.

Because the checks are now done outside normal request, we can have a
simpler failure strategy:

The check is now performed in the background every
`circuitbreaker_check_interval`. Failures are logged in redis. The
failures are reset when the check succeeds. Per check we will try
`circuitbreaker_access_retries` times within
`circuitbreaker_storage_timeout` seconds.

When the number of failures exceeds
`circuitbreaker_failure_count_threshold`, we will block access to the
storage.

After `failure_reset_time` of no checks, we will clear the stored
failures. This could happen when the process that performs the checks
is not running.

Fix https://gitlab.com/gitlab-org/gitlab-ce/issues/40295

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

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

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

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

Add `:broken_storage` metadata to examples to disable this behavior only
when necessary.

`migrate!` helper triggers a migration that is under the test, so we
need to revert it before each subsequent example. This means that we
need to place it in `before(:each, :migration)` hook.

We still want to migrate everything up *after* each *context*, so in
otherwords we need to place migrate up helper in `before(:context,
:migration)`.

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

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

So that we could make sure migration tests could run even if
geo is not setup in EE.

This is because we have a model like this:

``` ruby
class Geo::BaseRegistry < ActiveRecord::Base
  def self.connection
    raise 'Geo secondary database is not configured' unless Gitlab::Geo.geo_database_configured?

    super
  end
end
```

Also make it easier to have mailer helper

This was migrated to separate columns in 9.4, and now just needs to be removed
for real.

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