Don't fail builds for projects that are deleted when they are stuck
## What does this MR do?
Solves when dropping stuck connection.
```
Dropping stuck pending build 1545510 for runner
NoMethodError: undefined method `origin_merge_requests' for nil:NilClass
from /opt/gitlab/embedded/service/gitlab-rails/app/services/merge_requests/base_service.rb:50:in `merge_request_from'
from /opt/gitlab/embedded/service/gitlab-rails/app/services/merge_requests/base_service.rb:57:in `each_merge_request'
from /opt/gitlab/embedded/service/gitlab-rails/app/services/merge_requests/add_todo_when_build_fails_service.rb:5:in `execute'
from /opt/gitlab/embedded/service/gitlab-rails/lib/gitlab/metrics/instrumentation.rb:164:in `execute'
from /opt/gitlab/embedded/service/gitlab-rails/app/models/commit_status.rb:51:in `block (2 levels) in <class:CommitStatus>'
from /opt/gitlab/embedded/service/gem/ruby/2.1.0/gems/state_machines-0.4.0/lib/state_machines/eval_helpers.rb:79:in `call'
...
```
This happens, because that `default_scope` of `Projects` filters returns the projects that are not deleted, where `Ci::Build` doesn't take into account.
See merge request !4609
EmailsOnPushWorker was using the default queue, which made it impossible
to prioritize MergeWorker and other key workers with a separate Sidekiq
process.
The repository push email can be very expensive to generate, especially
with syntax-highlighted diffs. Instead of generating the email for each
recipient, generate one email object and reset the Message-Id and To
headers for each recipient. (Cloning would also be expensive in the case
of large emails, although probably not as bad as generating from
scratch.)
Added System Hooks for push and tag_push
Implements `push` and `tag_push` events for System Hooks.
They are based on Webhooks Events but without any deprecated item. We don't send commits too.
We are implementing this, to be used by Geo repository syncing (gitlab-org/gitlab-ee#76).
These hook events will be sent only when the respective configuration flags are set to true.
UI changes to admin screen will be made in another MR.
See merge request !3744
This caches the output of Repository#exists? in Redis while making sure
it's flushed properly when creating new repositories, deleting them,
etc.
For the ProjectWiki tests to work I had to make ProjectWiki#create_repo!
public as testing private methods in RSpec is a bit of a pain.
Reduce code complexity on GitPushService#execute
Code complexity for gitlab-ce after this has been refactored:
```
27.3: GitPushService#execute
```
This still needs to be merged into `gitlab-ee` presumably with conflicts... Perhaps we should create another issue for doing that?
I left the code sort of similar to what it was... If I could, I would have refactored most of the code into separate classes, etc. as it breaks probably all SOLID principles.
Fixes https://gitlab.com/gitlab-org/gitlab-ce/issues/13327
See merge request !2784
This ensures the caches for Repository#empty? and
Repository#has_visible_content? are flushed after a repository has been
imported or forked.
Fixesgitlab-org/gitlab-ce#13505
This removes the need for Sidekiq and any overhead/problems introduced
by TCP. There are a few things to take into account:
1. When writing data to InfluxDB you may still get an error if the
server becomes unavailable during the write. Because of this we're
catching all exceptions and just ignore them (for now).
2. Writing via UDP apparently requires the timestamp to be in
nanoseconds. Without this data either isn't written properly.
3. Due to the restrictions on UDP buffer sizes we're writing metrics one
by one, instead of writing all of them at once.
Stan Hu
Kamil TrzcinskiandPhil Hughes
James LopezandRobert Speicher
James LopezandRobert Speicher
Douwe Maan
Rubén Dávila
Sean McGivern
Jacob Vosmaer
Dmitriy Zaporozhets
Gabriel MazettoandDmitriy Zaporozhets
Zeger-Jan van de Weg
Yorick Peterse
Robert Speicher
Josh Frye
Douglas Barbosa Alexandre