Amplitude Status

Monitoring: Session replay should be processing as normal since 7:35 AM PT. We're keeping this incident open while we figure out how to recover the session replays sent during the incident. Efforts for that are underway. We'll provide the next update once we have a better sense of the timeline there.

Monitoring: Session replay since 7:30 AM has been stable.

We're still looking into the root cause of the issue between 4:55 AM PT - 7:30 AM PT and how to recover non-processed data.

Identified: We have switched to a new underlying queuing cluster underneath as of ~7:31 AM PT. So customers at this time shouldn't be seeing any new 5xxs on the session replay tracking endpoints, and the processing for these new session replays should be back to normal.

There are session replays sent between the period of 4:55 AM PT - 7:30 AM PT that might be missing or not processed yet. We're working to recover as many of them as possible.

Heatmaps were also affected as they rely on the same session replay ingestion infrastructure.

Investigating: Our session replay servers started to reject ~10% of the session replays being sent to Amplitude starting 4:55AM PT Aug 6 2025. All of our customers in the US cluster are impacted. We are also experiencing lags while processing the session replays at this time.

We are trying to recover our underlying clusters or failing over to a new one. Expect a new update in the next hour.

Investigating: Since 2025-08-06 04:55 PT (11:55 AM UTC), we are aware of an issue in session replay ingestion in the US cluster EU should be unaffected. Amplitude users may notice an absence of session replay data from the past 1 hour. Our team is actively investigating the source of the issue and will work to deploy a fix. There is likely some data loss; we're working to identify the issue and remediate it at this time. We will provide the next update in an hour.

Monitoring: Session replay should be processing as normal since 7:35 AM PT. We're keeping this incident open while we figure out how to recover the session replays sent during the incident. Efforts for that are underway. We'll provide the next update once we have a better sense of the timeline there.

Monitoring: Session replay since 7:30 AM has been stable.

We're still looking into the root cause of the issue between 4:55 AM PT - 7:30 AM PT and how to recover non-processed data.

Identified: We have switched to a new underlying queuing cluster underneath as of ~7:31 AM PT. So customers at this time shouldn't be seeing any new 5xxs on the session replay tracking endpoints, and the processing for these new session replays should be back to normal.

There are session replays sent between the period of 4:55 AM PT - 7:30 AM PT that might be missing or not processed yet. We're working to recover as many of them as possible.

Heatmaps were also affected as they rely on the same session replay ingestion infrastructure.

Investigating: Our session replay servers started to reject ~10% of the session replays being sent to Amplitude starting 4:55AM PT Aug 6 2025. All of our customers in the US cluster are impacted. We are also experiencing lags while processing the session replays at this time.

We are trying to recover our underlying clusters or failing over to a new one. Expect a new update in the next hour.

Investigating: Since 2025-08-06 04:55 PT (11:55 AM UTC), we are aware of an issue in session replay ingestion in the US cluster EU should be unaffected. Amplitude users may notice an absence of session replay data from the past 1 hour. Our team is actively investigating the source of the issue and will work to deploy a fix. There is likely some data loss; we're working to identify the issue and remediate it at this time. We will provide the next update in an hour.