Elasticsearch outage: EIS elevated 5xx error rates for model google-gemini-embedding-001

Affected components

• Elastic Inference Service

1. resolvedApr 20, 2026, 10:41 UTC

   An incident with an upstream service provider has been resolved, and access to the model is now restored.

2. investigatingApr 20, 2026, 09:50 UTC

   Errors are coming from the provider API (Google). The incident has been escalated to them, and we are still waiting for a response.

3. investigatingApr 20, 2026, 06:37 UTC

   We're seeing elevated 5xx error rates for the Gemini Embedding v1 model in EIS. The following default inference endpoint is affected: - .google-gemini-embedding-001 We're investigating and will update again in 2 hours or if there's a change in status.

What APIStatus.watch saw

Our 5-minute checks didn't record a change in Elasticsearch's overall status around this incident. Smaller or regional incidents often leave a provider's overall status green.

Other services with incidents at the same time

Overlapping incidents aren't necessarily related.

• 1Password: SSO sign-in failing on some attempts for some usersMajor · Apr 13, 2026, 17:24 UTC
• MongoDB Atlas: MongoDB Atlas BI Connector: Connections and queries failMajor · Apr 17, 2026, 17:03 UTC
• Snyk: CLI Test and Monitor DegradationMinor · Apr 17, 2026, 20:21 UTC

How often does Elasticsearch have incidents?

Elasticsearch reported 23 incidents in the last 90 days, 20 of them major or critical. A typical incident lasted 3h 35m. See Elasticsearch's uptime and incident history