Amplitude Incident - Session replay ingestion - 6 August 2025

All Session Replay services are now fully operational. The system is stable and functioning as expected from a customer perspective. While additional improvements will be implemented in the coming weeks to further harden the ingestion system, the immediate incident has been resolved.

Session replay ingestion is back to normal processing, but recovery efforts for replays during the incident are still underway.

Session replay and heatmaps were affected by an issue from 4:55 AM PT to 7:30 AM PT, but have been stable since then. The root cause is still under investigation.

We switched to a new underlying queuing cluster at ~7:31 AM PT, which should resolve 5xx errors on session replay tracking endpoints and restore normal processing of new session replays. However, session replays sent between 4:55 AM PT - 7:30 AM PT may be missing or unprocessed. Heatmaps were also affected as they rely on the same infrastructure.

We are investigating session replay ingestion issues causing ~10% rejection rates and processing lags in the US cluster.

Amplitude users may experience missing session replay data due to an ongoing ingestion issue.