Track and act upon the number of executed queriesquery-counts

This ensures that we have more visibility in the number of SQL queries
that are executed in web requests. The current threshold is hardcoded to
100 as we will rarely (maybe once or twice) change it.

In production and development we use Sentry if enabled, in the test
environment we raise an error. This feature is also only enabled in
production/staging when running on GitLab.com as it's not very useful to
other users.

2 files changed, 66 insertions, 0 deletions

diff --git a/doc/development/README.md b/doc/development/README.md
index 12cca9f84b7..45e9565f9a7 100644
--- a/doc/development/README.md
+++ b/doc/development/README.md
@@ -75,6 +75,7 @@ comments: false
 - [Ordering table columns](ordering_table_columns.md)
 - [Verifying database capabilities](verifying_database_capabilities.md)
 - [Database Debugging and Troubleshooting](database_debugging.md)
+- [Query Count Limits](query_count_limits.md)
 
 ## Testing guides
 
diff --git a/doc/development/query_count_limits.md b/doc/development/query_count_limits.md
new file mode 100644
index 00000000000..ebb6e0c2dac
--- /dev/null
+++ b/doc/development/query_count_limits.md
@@ -0,0 +1,65 @@
+# Query Count Limits
+
+Each controller or API endpoint is allowed to execute up to 100 SQL queries. In
+a production environment we'll only log an error in case this threshold is
+exceeded, but in a test environment we'll raise an error instead.
+
+## Solving Failing Tests
+
+When a test fails because it executes more than 100 SQL queries there are two
+solutions to this problem:
+
+1. Reduce the number of SQL queries that are executed.
+2. Whitelist the controller or API endpoint.
+
+You should only resort to whitelisting when an existing controller or endpoint
+is to blame as in this case reducing the number of SQL queries can take a lot of
+effort. Newly added controllers and endpoints are not allowed to execute more
+than 100 SQL queries and no exceptions will be made for this rule. _If_ a large
+number of SQL queries is necessary to perform certain work it's best to have
+this work performed by Sidekiq instead of doing this directly in a web request.
+
+## Whitelisting
+
+In the event that you _have_ to whitelist a controller you'll first need to
+create an issue. This issue should (preferably in the title) mention the
+controller or endpoint and include the appropriate labels (`database`,
+`performance`, and at least a team specific label such as `Discussion`).
+
+Once the issue has been created you can whitelist the code in question. For
+Rails controllers it's best to create a `before_action` hook that runs as early
+as possible. The called method in turn should call
+`Gitlab::QueryLimiting.whitelist('issue URL here')`. For example:
+
+```ruby
+class MyController < ApplicationController
+  before_action :whitelist_query_limiting, only: [:show]
+
+  def index
+    # ...
+  end
+
+  def show
+    # ...
+  end
+
+  def whitelist_query_limiting
+    Gitlab::QueryLimiting.whitelist('https://gitlab.com/gitlab-org/...')
+  end
+end
+```
+
+By using a `before_action` you don't have to modify the controller method in
+question, reducing the likelihood of merge conflicts.
+
+For Grape API endpoints there unfortunately is not a reliable way of running a
+hook before a specific endpoint. This means that you have to add the whitelist
+call directly into the endpoint like so:
+
+```ruby
+get '/projects/:id/foo' do
+  Gitlab::QueryLimiting.whitelist('...')
+
+  # ...
+end
+```