FeatureFlags

Not logged in - Log In / Register

Contents

  1. Feature Flags
    1. Key points
    2. Scenarios
    3. Concepts
    4. Priority
    5. Operations
    6. Web interface
    7. Debugging
    8. Performance
    9. Naming conventions
    10. Reading a feature flag
    11. Boolean values
    12. Adding and documenting a new feature flag
    13. Adding a new scope controller
    14. Testing
    15. Tips and traps
    16. See also

Feature Flags

FeatureFlags allow Launchpad's configuration to be changed while it's running, and for particular features or behaviours to be exposed to only a subset of users or requests.

Please note, that any changes to feature flags need to be recorded on https://wiki.canonical.com/InformationInfrastructure/OSA/LaunchpadProductionStatus

Key points

Scenarios

Concepts

A feature flag has a string name, and has a dynamically-determined value within a particular context such as a web or api request. The value in that context depends on determining which scopes are relevant to the context, and what rules exist for that flag and scopes. The rules are totally ordered and the highest-prority rule determines the relevant value.

Flags values are strings; or if no value is specified, None. (If an empty value is specified, the flag's value is the empty string).

For a list of available flags and scopes see https://launchpad.net/+feature-info

Priority

Prority is exposed as an integer that gives a total order across all rules for a particular flag. The numerically highest priority wins. For example with these rules 

hard_timeout team:admins 1 18000
hard_timeout default 0 15000

the first rule has a higher priority (1 > 0). So that rule is evaluated first, and it will match for anyone in ~admins. If that doesn't match, the second is evaluated and it is always true. So admins get a 18s timeout, and everyone else 15s.

Operations

A change to a flag in production counts as a production change: it is made by IS on request. Make the change on the appropriate wiki page (sorry, company internal), including an approval per the usual policy, and then ask in ~launchpad or #launchpad-dev.

Feature rules are loosely coupled to code changes: you can activate rules before the code they control is live.

Web interface

Debugging

A html comment at the bottom of rendered pages describes which features were looked up, and which scopes were consulted to make that decision. This doesn't include features that could be active but aren't relevant to the page, or scopes that may be active but aren't relevant to deciding the value of the features.

Performance

Feature flags are designed and intended to be fast enough that they can be used as much as is useful within reason. The result of a flag and of a scope is checked at most once per request.

If the page does not check any flags, no extra work will be done. The first time a page checks a flag, all the rules will be read from the database and held in memory for the duration of the request.

Scopes may be expensive in some cases, such as checking group membership. Whether a scope is active or not is looked up the first time it's needed within a particular request.

Naming conventions

Flag naming

Flags should be named as

where each of the parts is a legal Python name (so use underscores to join words, not dashes.)

The area is the general area of Launchpad this relates to: eg 'code', 'librarian', ...

The feature is the particular thing that's being controlled, such as 'recipes' or 'render_time'.

The effect is typically 'enabled', 'visible', or 'timeout'. These should always be in the positive sense, not 'disabled'. If timeouts are given, they should be in seconds (decimals can be given in the value.)

Scope naming

Scopes are matched using a simple regexp and for those that take parameters they are separated by a colon, eg team:admins.

There is no way at present to give a rule that checks multiple scopes or any other boolean conditions. You need to either choose one to match first, or add a new scope that matches just what you need, or extend the feature flag infrastructure to evaluate boolean expressions.

Reading a feature flag

(Note: "features/name" may not work! If you get a KeyError for "features," try "request/features/name instead.)

You can conditionally show some text like this 

  <tal:survey condition="features/user_survey.enabled">
    &nbsp;&bull;&nbsp;
    <a href="http://survey.example.com/">Take our survey!</a>
  </tal:survey>

You can use the built-in TAL feature of prepending not: to the condition, and for flags that have a value you could use them in tal:replace or tal:attributes.

If you just want to simply insert some text taken from a feature, say something like 

  Message of the day: ${motd.text}

Templates can also check whether the request is in a particular scope, but before using this consider whether the code will always be bound to that scope or whether it would be more correct to define a new feature: 

  <p tal:condition="feature_scopes/server.staging">
    Staging server: all data will be discarded daily!</p>

Boolean values

Frequently it is desired to have a boolean feature flag that can be used to toggle something on or off.

Decide what the default should be with the flag unset and this should be the False value of the boolean, so name the flag accordingly.

Then when checking the value do a bool() of the return value and use that as the value of the flag.

This means that unset and the empty string are False and anything else is True (note that this means that "false", "False", "off", 0, etc. all mean True)

For example 

    if getFeatureFlag('soyuz.frobble_the_wotsits.enabled'):
        wotsit.frobble()

Adding and documenting a new feature flag

If you introduce a new feature flag, as well as reading it from whereever is useful, you should also: 

# This table of flag name, value domain, and prose documentation is used to
# generate the web-visible feature flag documentation.
flag_info = sorted([
    ('code.recipes_enabled',
     'boolean',
     'enable recipes',
     ''),

The last item in that list is descriptive, not prescriptive: it documents the code's default behavior if no value is specified. The flag's value will still read as None if no value is specified, and setting it to an empty value still returns the empty string.

Adding a new scope controller

Add a new class in lib/lp/services/features/scopes.py and make sure it's in HANDLERS in that file. (You'll normally do this by adding it to WEBAPP_SCOPE_HANDLERS and/or SCRIPT_SCOPE_HANDLERS depending on whether it applies to webapp requests, scripts, or both).

Testing

FeatureFixture uses the testtools fixtures API to hook into your code. When it is installed on a TestCase object, the fixture will be automatically torn down and reset between tests, restoring all of the originally set flags.

NOTE: There is one gotcha: all existing flags are wiped out by the fixture for the duration of the test. If you want them to carry over, you need to do so yourself.

You can use the fixture three different ways:

  1. With the TestCase.useFixture() method

  2. As a context manager, using the 'with' statement

  3. By directly calling a fixture instance's setUp() and cleanUp() methods

Here is some sample code demonstrating each: 

   1   from lp.services.features.testing import FeatureFixture
   2   from lp.services.features import getFeatureFlag
   3 
   4 
   5   class FeatureTestCase(TestCase):
   6 
   7     layer = DatabaseFunctionalLayer  # Features need the database for now
   8 
   9     def test_useFixture(self):
  10       # You can use the fixture with the useFixture() TestCase method:
  11       self.useFixture(FeatureFixture({'reds': 'on'}))
  12       self.assertEqual('on', getFeatureFlag('reds'))
  13 
  14     def test_with_context_manager(self):
  15       # Or as a context manager:
  16       with FeatureFixture({'blues': None}):
  17         self.assertEqual(None, getFeatureFlag('blues'))
  18 
  19     def test_setUp_and_cleanUp(self):
  20       # You can call a fixture's setUp() and cleanUp() directly.
  21       # This is good for use in doctests.
  22       flags = FeatureFixture({'greens': 'mu'})
  23       flags.setUp()
  24       self.addCleanup(flags.cleanUp) # or use a try/finally

For more details on using the fixture and other feature flag utilities, check the module docs in lib/lp/services/features/__init__.py

For sample code, check:

Tips and traps

See also

FeatureFlags (last edited 2023-07-24 06:32:55 by jugmac00)