Experiment Management API Degradation

What to do

• Check whether you use Amplitude Management API. Check your stack

  Paste a package.json, lockfile, Terraform or a list of domains at down.guide/stack to see every provider you depend on.

• Confirm your retries, queues and scheduled jobs caught up after the 2h 29m it lasted.

  Anything that failed between 2026-03-20 17:31 UTC and 2026-03-20 20:00 UTC may need a replay.

• Subscribe to Amplitude's feed to hear about the next one first. RSS feed

  https://down.guide/feed?provider=amplitude works in Slack (/feed subscribe), Discord feed bots and any RSS reader.

Updates from Amplitude 3

1. identifiedMar 20, 2026, 5:31 PM UTC

   Since 2026-03-20 approximately 09:00 AM PT (16:00 UTC), the Experiment Flags Management API (/api/1/flags) has been returning incomplete results. Customers using the Flags Management API may see empty responses when listing feature flags, even though flags remain visible and functional in the Amplitude UI. The Experiments API (/api/1/experiments) is not affected. Our team has identified the root cause and a fix has been deployed. We are monitoring the rollout to confirm the issue is fully resolved. No data has been lost, and flags continue to evaluate correctly for end users. No customer action is needed at this time. We will provide the next update by 2026-03-20 12:00 PM PT (19:00 UTC).

   provider

2. monitoringMar 20, 2026, 7:13 PM UTC

   We have verified the fix, and the management-api is now fully operational. We will remain in a period of active monitoring to ensure continued stability.

   provider

3. resolvedMar 20, 2026, 8:00 PM UTC

   This incident has been resolved.

   provider

Corrections

None so far.

Something wrong on this page? Report an error or send us a note. Every fix is listed here.