Elastic Cloud Outage History

A recent change to Elastic Cloud's Privatelink implementation caused the deployment API to report incorrect URLs, resulting in connectivity issues for customers relying on those URLs. The incident lasted 23.2 hours and affected customers' ability to connect to their deployments through Privatelink. The issue was resolved by deploying a fix to the User Console that corrected the hostname reporting.

Elastic Cloud experienced a major 20.3-hour incident where users could not connect to Kibana through the Cloud UI via SAML SSO, specifically affecting PrivateLink customers while other hosted deployments remained unaffected. The engineering team identified the root cause, deployed a fix, and confirmed full resolution with no further customer impact.

The Elastic Cloud Synthetics service in the us-east-4 region experienced a major incident where customer monitor jobs failed to run on their expected schedules. The issue affected monitoring capabilities for customers using the Synthetics service in that region for approximately 2.8 hours. The problem was identified and resolved, restoring normal scheduling functionality for all affected monitor jobs.

Elastic Cloud experienced elevated 5xx error rates affecting Claude Sonnet 4.5 model inference endpoints, including chat completion and completion endpoints for both Anthropic Claude 4.5 Sonnet and GP-LLM-v2 models. The incident lasted 1.2 hours before all endpoints returned to normal operation.

Elastic Cloud experienced a major AutoOps outage in the AWS us-east-1 region lasting 2.9 hours, causing customer deployments to be incorrectly marked as inactive and preventing access to recent metrics. The engineering team identified the root cause, applied mitigations, and restored full AutoOps functionality. The incident was fully resolved after a monitoring period to ensure system stability.

Elastic Cloud experienced a major outage preventing customers from creating new Serverless projects in the GCP europe-west3 region for 6.9 hours. The engineering team identified the root cause and implemented a fix to restore project creation functionality. The issue was fully resolved and customers can now successfully create new Serverless projects in the affected region.

Elastic Cloud experienced a 9-hour service disruption in the GCP us-central1 region that impacted Kibana resource provisioning for new projects, the Docker registry availability, and Serverless project orchestration. Users encountered delays or failures when creating new projects, difficulties browsing and pulling Docker images, and slower provisioning and scaling of their hosted Serverless projects. The issue was resolved after working with hosting providers to restore the Docker registry and deployment orchestration services.

Elastic Cloud's fleet-server experienced errors that prevented new incoming connections for some projects, lasting 5.7 hours. The engineering team identified the root cause and successfully mitigated the issue. The incident was fully resolved after a monitoring period to ensure stability.