Amplitude change history

Amplitude reported a major degradation preventing delivery of automated exports that filter events across all export destinations, beginning 2026-09-22 23:50 UTC. The tracked incident lasted 6 h 5 min, from 2026-09-23 08:02 UTC to 2026-09-23 14:07 UTC. A fix restored all affected jobs; Amplitude expected no data loss and required no customer action.

Amplitude’s major Snowflake Data Share delivery outage began at 17:00 UTC (recorded start: 2026-09-22 18:49 UTC) and resolved at 2026-09-23 00:53 UTC; recorded duration: 6 h 4 min. Amplitude fixed an authentication issue; exports resumed normally, with no data loss expected or customer action needed, while existing shared data, ingestion and analytics remained unaffected.

Amplitude reported a critical outage affecting its UI, MCP and REST APIs, starting 2026-07-13 21:31 UTC and resolved at 2026-07-13 23:07 UTC (1 h 36 min). A fix was implemented; Amplitude expected no data loss, said the EU data center should be unaffected, and required no customer action.

Amplitude reported a major degradation affecting Agents and Chat: some users could not load or use AI-powered features. The incident started at 2026-07-13 18:51 UTC and was resolved at 2026-07-13 23:17 UTC, lasting 4 h 25 min; no cause or fix was stated.

Amplitude reported a major outage affecting the Custom Metrics metric details page, which could fail to load or display an error. The incident started at 2026-07-13 11:41 UTC and was resolved at 2026-07-13 14:59 UTC, lasting 3 h 17 min. A deployed fix restored normal loading; no data was lost.

Amplitude reported major elevated latency in the EU affecting Analytics - Web Application, Audiences - Web Application, Experiment - Web Application, Guides & Surveys, Session Replay and Web Application. The incident ran from 2026-07-13 06:15 UTC to 2026-07-13 08:39 UTC (2 h 23 min); a rollback restored performance.

Amplitude reported a major outage: shared embeds and public dashboard/chart links displayed blank pages between 00:24 UTC Jul 9 and 08:50 UTC. The incident started at 2026-07-09 09:12 UTC and resolved at 2026-07-09 09:45 UTC (33 min). Rolling back the affected service restored rendering; no data was lost and no customer action was needed.