Test Remote Config like production code. The app will read a resolved value by key, so the safest launch path is to make the server value easy to isolate and the client fallback easy to verify.

1. Start with a QA-safe surface

Use one of these patterns before broad rollout: AppActor does not currently expose a public SDK method for manually pinning one user to one Remote Config. Do not document or ship forced-user code for Remote Config testing.

2. Verify the SDK read path

For every key the app depends on, verify:
  • Fresh fetch returns the expected value.
  • Typed getters return null or nil when the type does not match.
  • Missing keys fall back to local app defaults.
  • JSON payloads are schema-checked before rendering.
  • Purchase and entitlement flows still use customer info.

3. Test cold start and cache behavior

Remote Config can load from memory or disk cache depending on platform and timing. Test:
  • first install with no cache
  • second launch with cached config
  • offline or transient network failure after a successful fetch
  • user identity changes if rules depend on entitlement state
  • app version changes when rules target version ranges

4. Launch gradually

Prefer this order for risky UI or purchase-surface changes:
  1. App-scoped QA config.
  2. Narrow version or country rule.
  3. Broader platform rule.
  4. Project-wide default after the payload is stable.

5. Roll back

The fastest rollback is usually one of:
  • deactivate the risky config
  • deactivate the risky rule
  • move the default value back to the previous production payload
  • lower traffic by moving the behavior into an experiment instead
Keep the previous payload shape compatible for at least one app release. That gives old clients a safe value while new clients roll out.

Next

Cache and Fallbacks

Learn what the SDK can return when the network is unavailable.

Launch A/B test from Remote Config

Move a stable config idea into measured experiment assignment.