Document features deployed behind feature flags

GitLab uses Feature Flags to strategically roll out the deployment of its own features. The way we document a feature behind a feature flag depends on its state (enabled or disabled). When the state changes, the developer who made the change must update the documentation accordingly.

Every feature introduced to the codebase, even if it’s behind a feature flag, must be documented. For context, see the latest merge request that updated this guideline.

Criteria

According to the process of deploying GitLab features behind feature flags:

  • By default, feature flags should be off.
  • Feature flags should remain in the codebase for a short period as possible to reduce the need for feature flag accounting.
  • In order to build a final release and present the feature for self-managed users, the feature flag should be at least defaulted to on.

See how to document them below, according to the state of the flag:

The **(FREE SELF)** badge or equivalent for the feature’s tier should be added to the line and heading that refers to enabling/disabling feature flags as Admin access is required to do so, therefore, it indicates that it cannot be done by regular users of GitLab.com.

Features disabled by default

For features disabled by default, add or improve the docs with every change in line with the definition of done.

Include details of the feature flag in the documentation:

  • Say that it’s disabled by default.
  • Say whether it’s enabled on GitLab.com.
  • If the feature can be enabled/disabled for a single project, add the by-project information. Otherwise, do not say anything about it.
  • Say whether it’s recommended for production use.
  • Document how to enable and disable it, preferably at the end of the file.
  • Add a warning to the user saying that the feature might be disabled.

For example, for a feature disabled by default, disabled on GitLab.com, cannot be enabled for a single project, and is not ready for production use:

# Feature Name

> - [Introduced](link-to-issue) in GitLab 12.0.
> - [Deployed behind a feature flag](<replace with path to>/user/feature_flags.md), disabled by default.
> - Disabled on GitLab.com.
> - Not recommended for production use.
> - To use in GitLab self-managed instances, ask a GitLab administrator to [enable it](#anchor-to-section). **(FREE SELF)**

This in-development feature might not be available for your use. There can be
[risks when enabling features still in development](<replace with path to>/user/feature_flags.md#risks-when-enabling-features-still-in-development).
Refer to this feature's version history for more details.

(...Regular content goes here...)

<!-- Add this at the end of the file -->

### Enable or disable <Feature Name> **(FREE SELF)**

<Feature Name> is under development and not ready for production use. It is
deployed behind a feature flag that is **disabled by default**.
[GitLab administrators with access to the GitLab Rails console](<replace with path to>/administration/feature_flags.md)
can enable it.

To enable it:

```ruby
Feature.enable(:<feature flag>)
```

To disable it:

```ruby
Feature.disable(:<feature flag>)
```

Adjust the blurb according to the state of the feature you’re documenting. Replace <Feature name>, **(FREE SELF)**, <feature flag>, and <replace with path to>, and #anchor-to-section accordingly.

Features that became enabled by default

For features that were released disabled by default but became enabled by default:

  • Say that it became enabled by default.
  • Say whether it’s enabled on GitLab.com.
  • If the feature can be enabled/disabled for a single project, add the by-project information. Otherwise, do not say anything about it.
  • Say whether it’s recommended for production use.
  • Document how to disable and enable it, preferably at the end of the file.
  • Add a warning to the user saying that the feature might be disabled.

For example, for a feature initially deployed disabled by default, that became enabled by default, that is enabled on GitLab.com, and is ready for production use:

# Feature Name

> - [Introduced](link-to-issue) in GitLab 12.0.
> - [Deployed behind a feature flag](<replace with path to>/user/feature_flags.md), disabled by default.
> - [Enabled by default](link-to-issue) in GitLab 12.1.
> - Enabled on GitLab.com.
> - Recommended for production use.
> - For GitLab self-managed instances, GitLab administrators can opt to [disable it](#anchor-to-section). **(FREE SELF)**

There can be
[risks when disabling released features](<replace with path to>/user/feature_flags.md#risks-when-disabling-released-features).
Refer to this feature's version history for more details.

(...Regular content goes here...)

<!-- Add this at the end of the file -->

### Enable or disable <Feature Name> **(FREE SELF)**

<Feature Name> is under development but ready for production use.
It is deployed behind a feature flag that is **enabled by default**.
[GitLab administrators with access to the GitLab Rails console](<replace with path to>/administration/feature_flags.md)
can opt to disable it.

To enable it:

```ruby
Feature.enable(:<feature flag>)
```

To disable it:

```ruby
Feature.disable(:<feature flag>)
```

Adjust the blurb according to the state of the feature you’re documenting. Replace <Feature name>, **(FREE SELF)**, <feature flag>, <replace with path to>, and #anchor-to-section accordingly.

Features directly enabled by default

For features enabled by default:

  • Say it’s enabled by default.
  • Say whether it’s enabled on GitLab.com.
  • If the feature can be enabled/disabled for a single project, add the by-project information. Otherwise, do not say anything about it.
  • Say whether it’s recommended for production use.
  • Document how to disable and enable it, preferably at the end of the file.
  • Add a warning to the user saying that the feature might be disabled.

For example, for a feature enabled by default, enabled on GitLab.com, that cannot be enabled for a single project, and is ready for production use:

# Feature Name

> - [Introduced](link-to-issue) in GitLab 12.0.
> - [Deployed behind a feature flag](<replace with path to>/user/feature_flags.md), enabled by default.
> - Enabled on GitLab.com.
> - Recommended for production use.
> - For GitLab self-managed instances, GitLab administrators can opt to [disable it](#anchor-to-section). **(FREE SELF)**

There can be
[risks when disabling released features](<replace with path to>/user/feature_flags.md#risks-when-disabling-released-features).
Refer to this feature's version history for more details.

(...Regular content goes here...)

<!-- Add this at the end of the file -->

### Enable or disable <Feature Name> **(FREE SELF)**

<Feature Name> is under development but ready for production use.
It is deployed behind a feature flag that is **enabled by default**.
[GitLab administrators with access to the GitLab Rails console](<replace with path to>/administration/feature_flags.md)
can opt to disable it.

To enable it:

```ruby
Feature.enable(:<feature flag>)
```

To disable it:

```ruby
Feature.disable(:<feature flag>)
```

Adjust the blurb according to the state of the feature you’re documenting. Replace <Feature name>, **(FREE SELF)**, <feature flag>, <replace with path to>, and #anchor-to-section accordingly.

Features enabled by project

If the feature can be enabled/disabled for a single project, include in the version history note:

> - It can be enabled or disabled for a single project.

Then add the by-project code to the code blocks:

Enable code:

# For the instance
Feature.enable(:<feature flag>)
# For a single project
Feature.enable(:<feature flag>, Project.find(<project id>))

Disable code:

# For the instance
Feature.disable(:<feature flag>)
# For a single project
Feature.disable(:<feature flag>, Project.find(<project id>))

For example, for a feature enabled by default, enabled on GitLab.com, that can be enabled by project, and is ready for production use:

# Feature Name

> - [Introduced](link-to-issue) in GitLab 12.0.
> - [Deployed behind a feature flag](<replace with path to>/user/feature_flags.md), enabled by default.
> - Enabled on GitLab.com.
> - Can be enabled or disabled for a single project.
> - Recommended for production use.
> - For GitLab self-managed instances, GitLab administrators can opt to [disable it](#anchor-to-section). **(FREE SELF)**

There can be
[risks when disabling released features](<replace with path to>/user/feature_flags.md#risks-when-disabling-released-features).
Refer to this feature's version history for more details.

(...Regular content goes here...)

<!-- Add this at the end of the file -->

### Enable or disable <Feature Name> **(FREE SELF)**

<Feature Name> is under development but ready for production use.
It is deployed behind a feature flag that is **enabled by default**.
[GitLab administrators with access to the GitLab Rails console](<replace with path to>/administration/feature_flags.md)
can opt to disable it.

To enable it:

```ruby
# For the instance
Feature.enable(:<feature flag>)
# For a single project
Feature.enable(:<feature flag>, Project.find(<project id>))
```

To disable it:

```ruby
# For the instance
Feature.disable(:<feature flag>)
# For a single project
Feature.disable(:<feature flag>, Project.find(<project id>))
```

Adjust the blurb according to the state of the feature you’re documenting. Replace <Feature name>, **(FREE SELF)**, <feature flag>, <replace with path to>, and #anchor-to-section accordingly.

Features with flag removed

Once the feature is ready and the flag has been removed, clean up the documentation. Remove the feature flag mention keeping only a note that mentions the flag in the version history notes:

# Feature Name

> - [Introduced](link-to-issue) in GitLab 12.0.
> - [Feature flag removed](link-to-issue) in GitLab 12.2.

(...Regular content...)