ES-2253532

We are actively investigating an issue affecting Databricks compute.

We will provide another update within 1 hour, or sooner if there is a material change in service status.

Impact Summary
Starting at 09:49 UTC on October 1, 2026, customers using Azure Databricks in West Europe may experience degraded availability with Serverless Compute. Affected customers may be unable to start or run serverless workloads for notebooks, workflows, and Lakeflow pipelines, and may encounter cluster allocation timeouts or errors.

Symptoms
• Serverless workloads failing to start
• Cluster allocation timeouts when launching serverless workloads
• Unexpected termination of serverless workloads

Recommendations
No customer action is required at this time.

Current Status
The issue is under active investigation.

Next Update
We will provide another update within 1 hour, or sooner if there is a material change in service status.

Impact Summary
Starting at 09:49 UTC on October 01, 2026, customers using Azure Databricks in West Europe may experience degraded availability with Serverless Compute, AI/BI, Databricks SQL, Declarative Pipelines, Jobs, and Notebooks. Affected customers may encounter workload start failures, cluster allocation timeouts, or errors when running queries, notebooks, jobs, and pipelines.

Symptoms
• Serverless workloads failing to start or encountering cluster allocation timeouts
• Jobs and pipeline runs failing, remaining pending, or not progressing
• SQL queries, dashboards, and notebook commands failing or timing out

Recommendations
No customer action is required at this time.

Current Status
The issue has been identified, and mitigation efforts are in progress in coordination with the cloud provider.

Next Update
We will provide another update within 1 hour, or sooner if there is a material change in service status.

Impact Summary
Starting at 09:49 UTC October 01, 2026, customers using Azure Databricks in West Europe may experience degraded availability or failures with Serverless Compute, AI/BI, Databricks Apps, Databricks SQL, Declarative Pipelines, Jobs, and Notebooks. Affected customers may encounter errors when starting workloads, cluster allocation timeouts, or failures when running queries, dashboards, and pipelines.

Symptoms
• Serverless workloads failing to start or encountering cluster allocation timeouts
• Jobs and pipeline runs remaining pending, failing, or not progressing
• SQL warehouses, AI/BI dashboards, and notebooks failing to start, load, or execute

Recommendations
No customer action is required at this time.

Current Status
The issue has been identified, and mitigation efforts are in progress in coordination with the cloud provider.

Next Update
We will provide another update within 1 hour, or sooner if there is a material change in service status.

Impact Summary
Starting at 09:49 UTC on October 01, 2026, customers using Azure Databricks in West Europe may experience failures or degraded availability with Serverless Compute, Databricks SQL (including Serverless SQL warehouses), AI/BI, Databricks Apps, Lakeflow Declarative Pipelines, Jobs, and Notebooks. Affected customers may see errors when starting workloads, cluster allocation timeouts, or unexpected workload terminations.

Symptoms
• Serverless workloads fail to start or hit cluster allocation timeouts.
• Jobs and pipeline runs stay pending, fail, or end unexpectedly.
• SQL warehouses, dashboards, notebooks, and apps fail to start, load, or run.

Recommendations
No customer action is required at this time.

Current Status
We have identified the cause. An issue with our cloud provider's infrastructure is affecting how virtual machines are provisioned for Azure Databricks in West Europe. Our engineering teams are working with the cloud provider to bring back provisioning capacity. We are also taking steps on our side to reduce load on the affected systems and help workloads recover. We are closely watching compute provisioning and service health while this work continues.

Next Update
We will provide another update within 1 hour, or sooner if service status changes significantly.