Google Cloud Outage History

Google Cloud Vertex AI Search for commerce customers experienced degraded performance with a 100% error rate observed for certain IPs. The issue affected multiple regions, including Global, EU, and US Vertex AI Search components, as well as the Global Recommendation AI component. The outage began on Feb 16, 2025 12:48 PM UTC (04:48 US/Pacific) and lasted a little over 3 hours, with recovery completed by Feb 16, 2025 3:55 PM UTC. Google Cloud issued three status updates during the incident, indicating ongoing investigation and gradual recovery. Affected customers likely encountered failures when performing Vertex AI Search queries and related recommendations during the window.

Google Cloud Chronicle Security feed-based ingestion began experiencing issues on Feb 13, 2025 8:13 PM UTC and primarily affected the Europe region, including Europe (regions) Chronicle Security and europe-west2 Chronicle Security. Users in Europe reported delays in detection and data ingestion, with symptoms centered on feed-based ingestion performance.

Chronicle Security users in the US-multiregions experienced an issue affecting the Risk Analytics Dashboard, with details not being up-to-date. The incident began at 2:00 AM UTC and lasted about 7 hours 55 minutes until 9:55 AM UTC. It affected the us / Chronicle Security and Multi-regions / Chronicle Security components.

Google Cloud engineering investigated and issued updates as the issue progressed, moving from warning to up status by the final update. The impact was geographically scoped to the US-multiregions region, with three official status updates published during the incident.

Google Cloud Chronicle Security WORKSPACE_ACTIVITY data ingestion in the asia-southeast1 and asia-south1 regions experienced delays, affecting users relying on workspace activity data in those regions. The incident began on Jan 31, 2025 12:53 AM UTC (Jan 30, 2025 16:53 US/Pacific) and was resolved by Jan 31, 2025 2:20 PM UTC, lasting approximately 13 hours and 27 minutes. Google Cloud issued four status updates during the incident as engineers investigated and worked to restore ingestion, with the two regional Chronicle Security components moving from degraded to normal as services recovered.

Google Cloud Translation experienced degraded performance with elevated latency and error rates, impacting translation requests across the globe. The incident began at 12:46 AM UTC and was resolved by 5:00 AM UTC, lasting roughly 4 hours. During the event, users may have encountered slower responses or intermittent service unavailability. The Cloud Translation component was listed in a degraded state and mitigation work was underway, with no ETA provided. The provider posted two status updates during the incident to communicate progress and the eventual recovery.

Google Cloud Translation services experienced degraded performance beginning at 12:50 PM UTC (04:50 US/Pacific), affecting the AutoML Translation and Cloud Translation components. Users likely saw elevated latency and translation errors, particularly for automated translation workflows. The incident lasted about 9 hours and 30 minutes, with ongoing mitigation activity and no initial ETA for resolution.

Over the course of the incident, Google Cloud issued three status updates. The updates progressed from noting the AutoML Translation issue to reporting elevated latency and error rates in Cloud Translation, and finally confirming a return to normal for Cloud Translation with the issue resolved by 10:20 PM UTC.

Google Cloud AppSheet experienced an outage affecting the us-east4 (Americas) and europe-west4 (Europe) regions. Customers were unable to load apps via the AppSheet editor or the app load page. The incident began at 6:45 PM UTC with regional AppSheet services degraded; us-east4 and europe-west4 were down. The outage lasted about 1 hour and 15 minutes and was resolved by 8:00 PM UTC. Google Cloud issued two status updates during the incident, noting that mitigation work was underway and providing no ETA at the time. User reports indicated loading issues and blank pages, with a noticeable spike around 6:00 PM UTC.

An outage affected Looker Studio, with dashboards and reports not loading across multiple regions. The disruption began at 2:07 PM UTC and was resolved by 2:56 PM UTC, lasting about 49 minutes. StatusGator detected the issue at Jan 24, 2025 2:07 PM UTC as an Early Warning Signal, with user reports confirming the problem before Google Cloud posted any official status updates. There were no Google Cloud status updates issued during the incident. The outage appeared globally, with reports from the United States, Brazil, Canada, and Mexico indicating Looker Studio was not loading or is unresponsive.

Chronicle Security on Google Cloud experienced elevated timeout errors across US regions, impacting views and APIs related to rules and detections, as well as several UI pages. The issue began on 8:30 PM UTC (12:30 US/Pacific) and affected two Chronicle Security components: Multi-regions and US region.

Google Cloud posted updates on the incident, with three status changes over its duration. The incident moved from up to warn at Jan 15, 2025 10:05 PM UTC and was later resolved by Jan 15, 2025 11:45 PM UTC. The outage lasted approximately 3 hours and 15 minutes, with users experiencing timeouts affecting views and APIs related to rules and detections and related UI pages across US regions, before the Chronicle Security components returned to normal.

Google Cloud's Vertex Gemini API (Gemini 1.5 Flash 002 model) began experiencing elevated errors on Jan 10, 2025 7:36 PM UTC (11:36 US/Pacific), with requests returning 5XX or 429 responses. The disruption affected Vertex Gemini API across seven components, spanning Europe (regions) / Vertex Gemini API, europe-west2, europe-west3, Asia Pacific (regions) / Vertex Gemini API, asia-northeast1, asia-south1, and australia-southeast1. The outage lasted about 66 hours and 44 minutes and was resolved by Jan 13, 2025 2:20 PM UTC.

The Google Cloud status page documented two official updates during the incident: an initial transition from up to warn at Jan 10, 2025 8:40 PM UTC as the issue was investigated, followed by a restoration to up at Jan 13, 2025 2:20 PM UTC as services returned to normal. Affected users experienced API request failures and degraded performance while the issue was active.