Enable and disable GitLab features deployed behind feature flags

DETAILS: Tier: Free, Premium, Ultimate Offering: Self-managed

GitLab adopted feature flags strategies to deploy features in an early stage of development so that they can be incrementally rolled out.

Before making them permanently available, features can be deployed behind flags for a number of reasons, such as:

To test the feature.
To get feedback from users and customers while in an early stage of the development of the feature.
To evaluate users adoption.
To evaluate how it impacts the performance of GitLab.
To build it in smaller pieces throughout releases.

Features behind flags can be gradually rolled out, typically:

The feature starts disabled by default.
The feature becomes enabled by default.
The feature flag is removed.

These features can be enabled and disabled to allow or prevent users from using them. It can be done by GitLab administrators with access to the Rails console or the Feature flags API.

When you disable a feature flag, the feature is hidden from users and all of the functionality is turned off. For example, data is not recorded and services do not run.

If you used a certain feature and identified a bug, a misbehavior, or an error, it's very important that you provide feedback to GitLab as soon as possible so we can improve or fix it while behind a flag. When you upgrade GitLab, the feature flag status may change.

Risks when enabling features still in development

Before enabling a disabled feature flag in a production GitLab environment, it is crucial to understand the potential risks involved.

WARNING: Data corruption, stability degradation, performance degradation, and security issues may occur if you enable a feature that's disabled by default.

Features that are disabled by default may change or be removed without notice in a future version of GitLab.

Features behind default-disabled feature flags are not recommended for use in a production environment and problems caused by using a default disabled features aren't covered by GitLab Support.

Security issues found in features that are disabled by default are patched in regular releases and do not follow our regular maintenance policy with regards to backporting the fix.

Risks when disabling released features

In most cases, the feature flag code is removed in a future version of GitLab. If and when that occurs, from that point onward you can't keep the feature in a disabled state.

How to enable and disable features behind flags

Each feature has its own flag that should be used to enable and disable it. The documentation of each feature behind a flag includes a section informing the status of the flag and the command to enable or disable it.

Start the GitLab Rails console

The first thing you must do to enable or disable a feature behind a flag is to start a session on GitLab Rails console.

For Linux package installations:

sudo gitlab-rails console

For installations from the source:

sudo -u git -H bundle exec rails console -e production

For details, see starting a Rails console session.

Enable or disable the feature

After the Rails console session has started, run the Feature.enable or Feature.disable commands accordingly. The specific flag can be found in the feature's documentation itself.

To enable a feature, run:

Feature.enable(:<feature flag>)

Example, to enable a fictional feature flag named example_feature:

Feature.enable(:example_feature)

To disable a feature, run:

Feature.disable(:<feature flag>)

Example, to disable a fictional feature flag named example_feature:

Feature.disable(:example_feature)

Some feature flags can be enabled or disabled on a per project basis:

Feature.enable(:<feature flag>, Project.find(<project id>))

For example, to enable the :example_feature feature flag for project 1234:

Feature.enable(:example_feature, Project.find(1234))

Some feature flags can be enabled or disabled on a per user basis. For example, to enable the :example_feature flag for user sidney_jones:

Feature.enable(:example_feature, User.find_by_username("sidney_jones"))

Feature.enable and Feature.disable always return true, even if the application doesn't use the flag:

irb(main):001:0> Feature.enable(:example_feature)
=> true

When the feature is ready, GitLab removes the feature flag, and the option for enabling and disabling it no longer exists. The feature becomes available in all instances.

Check if a feature flag is enabled

To check if a flag is enabled or disabled, use Feature.enabled? or Feature.disabled?. For example, for a feature flag named example_feature that is already enabled:

Feature.enabled?(:example_feature)
=> true
Feature.disabled?(:example_feature)
=> false

When the feature is ready, GitLab removes the feature flag, and the option for enabling and disabling it no longer exists. The feature becomes available in all instances.

View set feature flags

You can view all GitLab administrator set feature flags:

Feature.all
=> [#<Flipper::Feature:198220 name="example_feature", state=:on, enabled_gate_names=[:boolean], adapter=:memoizable>]

# Nice output
Feature.all.map {|f| [f.name, f.state]}

Unset feature flag

You can unset a feature flag so that GitLab falls back to the current defaults for that flag:

Feature.remove(:example_feature)
=> true