On March 27, 2026, between 10:56 and 12:50 UTC, the dbt Platform experienced a service disruption in the Azure US MC environment. During this 114-minute window, customers were unable to access the dbt Platform, trigger jobs, or run scheduled tasks in this environment.
During a routine proactive database scaling operation, a software defect in Azure's scaling workflow combined with a compute capacity shortage in the region caused the production database to become unavailable. Our engineering team immediately escalated to Azure support, who were able to mitigate the issue and restore database availability. No other environments were affected, and no data loss or data corruption occurred as a result of this incident.
Duration: 114 minutes (10:56–12:50 UTC)
Affected environment: Azure US MC only; no other regions or environments were impacted
Job execution: During the outage window, customers were unable to trigger new jobs. Customers also experienced scheduled and in-progress jobs being cancelled, delayed, or failing. After recovery, normal job scheduling resumed.
Data integrity: No data loss or data integrity issues occurred. The database was fully intact upon recovery.
During a planned infrastructure maintenance operation, a software defect in our cloud provider’s internal systems caused a failure that left our production database unavailable. Specifically, Azure's systems allowed a database scaling operation to proceed despite not having sufficient capacity to complete it and when the operation failed, it did not automatically restore the database to its previous state.
Recovery required manual intervention by Azure's engineering team, which extended the duration of the outage.
Our engineering team detected the outage within minutes and immediately opened a high-severity support case with Microsoft Azure. Azure escalated to their on-call engineering team, who diagnosed the stalled scaling workflow and restored database availability. The 114-minute duration was largely driven by the time required for Azure's engineering team to be engaged and complete the manual recovery. All dbt Platform services recovered shortly after database availability was restored.
Azure-side fixes:
Microsoft Azure has patched the software defect in their pre-flight validation and updated their capacity checks for the affected region, ensuring that future scaling requests are rejected immediately when resources are unavailable, before any changes are made to the running database.
dbt Labs-side improvements:
We have introduced additional pre-scaling safeguards and readiness checks before initiating scaling operations, reducing our reliance on Azure's pre-flight checks alone.
While the region remains under capacity constraints, we are coordinating all scaling operations directly with Azure's engineering team in real time, ensuring immediate support is available should any issues arise.
We are evaluating changes to our scaling procedures to ensure that database availability is preserved in the event of a failed operation, independent of vendor rollback behavior.
If you have any questions about this incident or its impact on your account, please reach out to your dbt Labs account team or contact dbt Labs Support.
