Header Image
All Systems Operational

Updated a few seconds ago

Welcome to Azure Databricks status page. Here you will find high-level availability information for all of the Azure Databricks services.

AI/BI

Operational


Compute

Operational


Databricks Apps

Operational


Databricks SQL

Operational


Lakebase

Operational


Lakeflow

Operational


Mosaic AI

Operational


Notebooks

Operational


Unity Catalog

Operational


Scheduled Maintenance

Emergency Maintenance Planned Maintenance

Schedule

January 23, 2026 17:00 - January 24, 2026 17:00 UTC

Components

AI/BI, Mosaic AI

Locations

Germany West Central, Norway East, Sweden Central, Switzerland North, Switzerland West, West Europe, North Europe, UK West, UK South, South Africa North, UAE North, France Central, Qatar Central

Description

An emergency cloud-provider maintenance event will affect some GPU-accelerated compute resources. Maintenance Window - Start: Friday, January 23, 2026, at 17:00 UTC - End: Saturday, January 24, 2026, at 17:00 UTC (Expected) Services Impacted During this maintenance period, customers may experience interruptions or downtime across the following Databricks services: - Open Source Foundation Model Serving (Llama, Qwen, Gemma, GPT OSS, GTE, BGE) - AI Functions (including ai_query and ai_parse_document) - Vector Search - Databricks Assistant - Databricks Data Classification Workloads actively running on GPU instances at the start of the maintenance window may be interrupted, placed in a queue, or require a manual restart once the window concludes. We advise completing or re-routing time-sensitive GPU tasks prior to the scheduled start time if possible. Action for Mitigation: Enable Cross-Geo Processing To minimize service disruptions, we strongly recommend that customers enable cross-Geo processing where possible. Refer to: https://learn.microsoft.com/en-us/azure/databricks/resources/databricks-geos#cross-geo-processing ​​Note 1: Enabling cross-Geo processing may result in workload execution or data processing occurring outside of your primary region. For customers with data residency, data locality, or regulatory constraints, this may not be permissible. Please review your organizational policies and compliance requirements before enabling this setting. Note 2: Llama 3 Provisioned Throughput will remain impacted by this event regardless of whether cross-Geo processing is enabled. If you are using Llama 3 Provisioned Throughput, we recommend upgrading to a newer model, such as GPT-OSS, and enabling cross-geo processing.