Gemini Major Outage — May 5, 2026

Tickerr probes detect first overloaded (HTTP 503) responses from gemini-2.5-flash. Both parallel probes return errors. gemini-2.5-flash-lite responds but with elevated TTFT of 3,375–6,846ms (vs baseline ~900ms).

Consecutive failures confirmed across all 5-minute checks. gemini-2.5-flash returning overloaded on every probe. gemini-2.5-flash-lite remains up but degraded. Tickerr opens a synthetic incident.

Outage continues. 15+ consecutive overloaded checks recorded on gemini-2.5-flash. Flash-lite TTFT at 939–1,927ms — still elevated. No acknowledgment from Google.

First signs of partial recovery: one probe on gemini-2.5-flash succeeds with TTFT 1,490ms. Other probe still returns overloaded. Flash-lite TTFT 689–1,431ms.

gemini-2.5-flash recovering: TTFT drops to 792–3,193ms range. Flash-lite TTFT elevated at 7,124ms. Recovery is intermittent — additional overloaded errors recorded at 10:10 and 10:20–10:40 UTC.

gemini-2.5-flash nominally operational (664–1,038ms) but flash-lite continues severely degraded with TTFT 7,124–7,127ms. Intermittent overloaded errors on flash resume at 11:20 UTC. Elevated latency persists across both models throughout the afternoon.

Google publishes a status update acknowledging "Issues with streaming Deep Research" — resolved at 19:07 UTC. Google does not mention the API overloads affecting gemini-2.5-flash that Tickerr detected 11+ hours earlier. Flash-lite TTFT reaches 13,942ms at 19:00 UTC.

All models fully operational. gemini-2.5-flash TTFT normal at 548–1,520ms. gemini-2.5-flash-lite TTFT normal at 375ms. Total outage duration for flash: ~2h 45min of active API failures (08:01–10:50 UTC). Flash-lite degradation lasted 11+ hours (08:01–19:20 UTC). Community impact: 44 users reported issues on Tickerr during the incident.