Statsig adapter

[AAorris]

Statsig Adapter

Implements a Statsig adapter using statsig-node and statsig-node-vercel.

The adapter takes initialization details for Statsig and the Edge Config adapter, and returns an interface that can be used to declare Feature Gates, Feature Gates with metric lifts, Experiments, DynamicConfigs, and Autotunes.

The adapter will initialize Statsig, notably with defaults where IDLists are not synced and there is no strategy to sync them. This can be changed by passing alternatives in statsigOptions, but these defaults reduce network fetching when flags are precomputed in middleware.

The decide function will not log events to Statsig — Instead, it's expected that this will be done client side. The reasoning for this is that Next.js may prefetch routes, and we do not want to expose a user to experiments or gates just because they were prefetched. Additionally, because of the serverless runtime context, the buffering and flushing of logs can be unreliable.

For this reason, we will expect the client to log events, for which they will either need to log events manually with a rule ID, or take these results as initial values while checking them again with exposures on the client side.

Simple usage

The usage of calling the adapter directly is the simplest case — A feature gate without metric lifts.

const statsigAdapter = createStatsigAdapter({
  statsigSecretKey: 'statsigSecretKey',
});

export const exampleFlag = flag<boolean, StatsigUserEntities>({
  key: 'example',
  adapter: statsigAdapter(),
});

Methods

featureGate — Receives the boolean value and ruleID and can map the result to a flag value

export const featureGateFlag = flag<[boolean, string], StatsigUserEntities>({
  key: 'example-gate',
  adapter: statsigAdapter.featureGate((value, ruleId) => [value, ruleId]),
});

dynamicConfig — Receives the evaluated value of a DynamicConfig object and ruleId and can map the result to a flag value

export const dynamicConfigFlag = flag<string, StatsigUserEntities>({
  key: 'example-config',
  adapter: statsigAdapter.dynamicConfig((value) => value.heroText),
});

experiment — Receives the evaluated value of an Experiment object and ruleId and can map the result to a flag value

export const dynamicConfigFlag = flag<'control' | 'test', StatsigUserEntities>({
  key: 'example-experiment',
  adapter: statsigAdapter.experiment((value, ruleId) => value.experimentGroup),
});

autotune — Receives the evaluated value of an Autotune and ruleId and can map the result to a flag value

const experimentGroups = ['control', 'variant1', 'variant2'] as const;
type ExperimentGroup = (typeof experimentGroups)[number];

export const dynamicConfigFlag = flag<ExperimentGroup, StatsigUserEntities>({
  key: 'example-autotune',
  adapter: statsigAdapter.autotune((value, ruleId) => value.experimentGroup),
});

initialize — Statsig.initialize will be called lazily, when a flag's decide method is invoked. To initialize Statsig earlier using the adapter's implementation:

statsigAdapter.initialize()

[vercel]

The latest updates on your projects. Learn more about Vercel for Git ↗︎

Name	Status	Preview	Comments	Updated (UTC)
flags-sdk-snippets	✅ Ready (Inspect)	Visit Preview	💬 Add feedback	Jan 20, 2025 4:37pm
summer-sale	✅ Ready (Inspect)	Visit Preview	💬 Add feedback	Jan 20, 2025 4:37pm

[socket-security]

New and updated dependencies detected. Learn more about Socket for GitHub ↗︎

Package	New capabilities	Transitives	Size	Publisher
npm/[email protected] 🔁 npm/[email protected]	environment	0	238 kB	react-bot
npm/[email protected]	Transitive: network	+8	2.52 MB	kenny-statsig
npm/[email protected]	network	+7	2.63 MB	xinlili

View full report↗︎