Amplitude Outage History

This incident has been resolved.

The Amplitude web application in the US was briefly inaccessible. We identified the issue to be related to a misconfigured feature flag. The issue started at 10:00 AM PST, lasting 10 minutes, and is now resolved.

Apr 15, 18:11 PDT Resolved - This incident has been resolved. Apr 15, 17:21 PDT Update - Event streaming has largely recovered. Fewer than 1% of events are still experiencing delays of over one hour. We’ll provide another update once latency has returned to normal for all data. Apr 15, 17:20 PDT Update - Event streaming has largely recovered. Fewer than 1% of events are still experiencing delays of over 30 minutes. We’ll provide another update once latency has returned to normal for all data. Apr 15, 16:09 PDT Monitoring - Since 2026-04-15 12:20 PT (19:20 UTC), Amplitude has been experiencing high latency in Event Streaming. Current status: a) Our ingestion and data processing systems are working as expected and no data is lost. b) All of our customers are impacted. c) Impacted customers will see delayed event data in 3rd party Event Streaming destinations for the last 3 hrs. Our team has identified the issue and is actively working on a resolution. There is no data loss, and data will still be sent to downstream destinations. No customer action is needed at this time. We will provide the next update in an hour.

Apr 13, 14:00 PDT Resolved - This incident has been resolved. Apr 13, 13:53 PDT Monitoring - Erroneous events have been deleted. Event property limits should be accurate again and new event properties should be usable again. Any event properties that came in only during the overage window (and haven't been sent again since) will not be usable until post-processing occurs at end of day. Apr 13, 08:45 PDT Identified - At 8:45 am PDT, Guides and Surveys added a large number of event properties. Which projects are impacted? Any project that previously had less than 2000 event properties and that now has 2000 or more event properties. What's the impact? All such projects are unable to (a) add new event properties and (b) use those new event properties for filtering, creating cohorts. The issue has been identified, and the team is working on a fix that will remove all of these event properties. You may see some irregularities during this time. Once the mistakenly-added event properties are deleted, some post-processing will occur at the end of the day to restore any new event properties which could not be added to event property overages (new ones, added after 8:45 am PDT, assuming the project was put over the 2000 event property limit due to this change).

Apr 12, 03:52 PDT Resolved - This incident has been resolved. Apr 12, 03:47 PDT Monitoring - A fix has been implemented and we are monitoring the results. Apr 12, 02:52 PDT Update - We are continuing to investigate this issue Apr 12, 01:24 PDT Investigating - Since 2026-04-12 00:43 PT (07:43 UTC), the EU data collection endpoint (api.eu.amplitude.com) has been experiencing service disruptions. Customers sending event data to the EU endpoint may encounter errors, and data ingestion may be temporarily unavailable. Our team is actively investigating the source of the issue and working to restore normal operations. Auto-scaling measures have been initiated. We do not expect permanent data loss for events that are retried by client SDKs. No customer action is needed at this time beyond ensuring your SDKs are configured with retry logic. We will provide the next update by 2026-04-12 02:00 PT (09:00 UTC).

Apr 6, 17:48 PDT Resolved - This incident has been resolved. Apr 6, 16:39 PDT Monitoring - A fix has been implemented, and we are monitoring recovery. Some users may need to re-authenticate the following components to restore full functionality: • Ampli CLI • MCP Server • Notion and Atlassian chart previews Apr 6, 12:53 PDT Investigating - We are currently investigating an issue affecting user authentication for: Ampli CLI, MCP Server, Notion/Atlassian chart previews.

Apr 3, 16:52 PDT Resolved - This incident has been resolved. Apr 3, 16:33 PDT Update - All systems are operational. Apr 3, 15:23 PDT Update - Our data pipeline are catching up with no expected data loss, incident is expected to be fully resolved shortly. Apr 3, 15:14 PDT Monitoring - We're continuing to monitor as services recover. Apr 3, 15:07 PDT Update - We have identified the issue and a fix has been deployed, services are being restored. Analytics event collection is not been impacted and no data loss is expected. Apr 3, 14:56 PDT Update - We are continuing to investigate this issue. Apr 3, 14:52 PDT Investigating - We are currently investigating a site-wide outage.

Mar 27, 18:30 PDT Completed - The scheduled maintenance has been completed. Mar 27, 18:07 PDT Verifying - Configuration update has been complete; currently verifying. Mar 27, 17:00 PDT In progress - Scheduled maintenance is currently in progress. We will provide updates as necessary. Mar 26, 13:00 PDT Scheduled - We will be applying underlying configuration changes that require a short maintenance window. Some page loads may fail for a short time during this window.

Mar 27, 10:11 PDT Resolved - We have fixed the issue Mar 27, 09:42 PDT Identified - We have identified a fix. We are rolling the fix out. Mar 27, 08:52 PDT Investigating - For some flags, the exposures chart is not loading properly. For some experiments, the analysis page is not loading properly saying "This experiment's user property is not available yet. It will appear after users are assigned." and "Invalid user property gp:". We are working on a fix.

Some Flags and experiments that were targeting cohorts did not evaluate users into control and treatment correctly. This has been fixed

Mar 25, 11:09 PDT Resolved - This incident has been resolved. Mar 25, 11:00 PDT Investigating - Components Affected: Experiments Message: Since 2026-03-25 10:09 AM PT (17:09 UTC), users have been unable to create new experiments in Amplitude Experiment. Existing experiments and other Experiment functionality are not affected. Our team has identified the source of the issue and is actively working to deploy a fix. We expect no data loss. No customer action is needed at this time. We will provide the next update by 12:00 PM PT (19:00 UTC).

Mar 20, 13:00 PDT Resolved - This incident has been resolved. Mar 20, 12:13 PDT Monitoring - 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. Mar 20, 10:31 PDT Identified - 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).

Duration: March 16, 2026, 11:20 AM PT to 1:17 PM PT Between 11:20 AM PT and 1:17 PM PT, the Experiment Evaluation Service experienced a partial outage, resulting in elevated 500, 502, and 503 errors for some requests. Our team has successfully resolved the issue, and the service is now operating normally. We apologize for any inconvenience this disruption may have caused.

Between 2026-03-16 08:57 AM PT (15:57 UTC) and 09:23 AM PT (16:23 UTC), some Amplitude users experienced intermittent 500 errors when loading charts, dashboards, and reports in Analytics and Experiment. A second brief interruption occurred at approximately 09:49 AM PT (16:49 UTC), during which some requests again returned 500 errors. The issue was caused by a temporary database disruption that has since automatically recovered. Our team has confirmed that services are operating normally. No data loss occurred, and no customer action is needed. We are actively investigating the root cause to prevent recurrence and will continue to monitor the system.

A migration to transition some Growth orgs to a new permission system, which is activated by a feature flag, did not operate correctly, resulting in the appearance of lost role assignments for users. This affected some Growth orgs that have project-based permissions. Removing the feature flag for these orgs reverted them to the old permission system, restoring all role assignments correctly.

Mar 2, 20:46 PST Resolved - This incident has been resolved. Mar 2, 20:22 PST Monitoring - A fix has been implemented and we are monitoring the results. Mar 2, 18:53 PST Investigating - We've identified an issue with search queries failing within the Resource Center due to an issue with an upstream provider. We are working with them to identify the root cause.

Feb 27, 15:47 PST Resolved - As of 3:10 PM PT, we have fully recovered and all systems are operating normally. Feb 27, 14:22 PST Identified - Since 2026-02-27 1:14 PT, some Amplitude users may have experienced issues while loading charts/dashboards. Our team is actively investigating the source of the issue and working to deploy a fix. New data ingested recently might be delayed and not show up in results by up to 30 mins. We expect no data loss. No customer action is needed at this time. We will provide the next update by 2026-02-27 3:00 PM PT.

Feb 26, 14:58 PST Resolved - This incident has been resolved. Feb 26, 12:56 PST Monitoring - The rollback has been applied and we are monitoring recovery. Feb 26, 12:55 PST Identified - We've identified the root cause and rollback has been triggered. Feb 26, 12:48 PST Update - We are continuing to investigate this issue. Feb 26, 12:48 PST Investigating - Since 2026-02-26 12:00 PT (20:00 UTC), some Amplitude users may experience a "Something went wrong" error page when loading the Amplitude application. Users with certain permission configurations may be unable to access the platform. Our team is actively investigating the source of the issue and working to deploy a fix. We expect no data loss. No customer action is needed at this time. We will provide the next update by 2026-02-26 1:30 PM PT (21:30 UTC).

Feb 12, 15:49 PST Resolved - This incident has been resolved. The bug was isolated to a permissions issue with Amplitude users and did not impact any data ingestion, experiments, or other features. All user permissions should be back to operational. Feb 12, 15:31 PST Identified - The issue has been identified and an initial fix pushed. Feb 12, 15:19 PST Investigating - Since 2026-02-12 02:27 PT, we have been experiencing issues with user permissions in Amplitude. Currently, some users may not be able to access features they should have access to. Our team is actively investigating the source of the issue and will work to deploy a fix. We expect no data loss. No customer action is needed at this time.