ELASTIC

🟡 AutoOps got caught playing hide-and-seek in GCP us-central-1 and lost spectacularly for 42 minutes. Engineers found the body, applied a fix, and are now nervously refreshing dashboards to make sure it doesn't pull a Lazarus. (Auto-resolved: no updates in 24h)

⏰ Jan 19, 2026 Fixed (for now) spotted via status_rss

// The official story

AutoOps Partial Outage in GCP us-central-1

// Status page says

Dec 23 , 05:29 UTC Resolved - The system is operational. Dec 23 , 04:47 UTC Monitoring - We are seeing recovery in the system with the deployed fix, and are continuing to monitor the health of the system. Dec 23 , 04:24 UTC Identified - We have identified the underlying cause of the issue and are applying a fi

🔗 Official status page (for the PR-approved version)

https://status.elastic.co/incidents/cd0zfzw0kpn2

Elastic's Greatest Hits

🟢 ELASTIC 5d ago

🟡 Elastic's GCP us-central1 region decided projects should come with a 'read receipts disabled' feature. Some Serverless deployments are either moving at geological timescales or vanishing entirely. The status page says 'investigating' which is tech speak for 'we're refreshing our monitors and hoping it fixes itself.'. (Auto-resolved: no updates in 24h)

🟢 ELASTIC Sep 17, 2026

🟡 Elastic's AutoOps can't find its node metrics (CPU, memory, disk) in certain regions—they're just gone from recent time ranges. The good news: your data's fine, cluster health is fine, everything else is fine. It's just the one thing you need to see that's MIA. ETA on the scavenger hunt: 2 hours. (Auto-resolved: no updates in 24h)

🟢 ELASTIC Sep 16, 2026

🟡 Elastic's own support portal decided to become unsupported. The company that helps you monitor everything apparently turned off monitoring for, well, everything. They're 'investigating' with their provider—translation: playing hot potato with blame while customers scream into the void. (Auto-resolved: no updates in 24h)

⚪ ELASTIC Aug 27, 2026

⚪ Someone deployed a DNS config change that yeet'd .apm endpoints into the void. Serverless customers got the silent treatment for ~9 hours until engineers remembered these endpoints existed. Elastic's internal DNS provider apparently needed a nap.