Target lists with millions of IDs using Remote Saved Groups
Remote Saved Groups let you target very large lists, such as millions of user or account IDs. GrowthBook keeps the IDs outside the SDK payload, so you can target them in feature and experiment rules without growing what each client downloads.
Why ID Lists stop scaling
An ID List Saved Group travels inside the SDK payload, and each list is limited to 1 MB. Large lists make the payload bigger for every client that fetches it.
Use Remote Saved Groups for lists too large for an ID List. You still target with the same Saved Groups rule or the is in the saved group and is not in the saved group conditions.
How lookups work
- You upload a CSV of IDs, up to 100 MB by default, or roughly 3 to 5 million IDs. You can also keep the IDs in your own storage.
- Each upload creates a new version. It changes targeting only after you add it to a revision and publish it through your normal review flow.
- With remote evaluation, GrowthBook Proxy 1.6.0 or later loads the IDs into Redis and looks up each user's groups before evaluating. Your app needs no changes, and the Proxy ignores any group IDs the client sends.
- Without the Proxy, the JavaScript SDK 2.0.0 or later can run the lookup with a resolver. Other SDKs can set the user's groups after your own lookup.
If the Proxy can't check a user's group membership, it skips rules that depend on that remote group, including rules that exclude the group. That can happen if the group hasn't loaded or if the lookup times out.
Getting started
Remote Saved Groups are an Enterprise feature. Go to SDK Configuration → Saved Groups, open the Remote Groups tab, and click Add Remote Group. Choose the attribute the IDs belong to, which must be a string or number, then click Upload CSV on the group's page. Self-hosted GrowthBook needs UPLOAD_METHOD set to s3 or google-cloud to accept uploads.
The Remote Groups documentation covers Proxy setup, custom lookups, and the REST API upload flow.
