We're seeing intermittent Azure Storage failures across several customer-owned tenants in West US 2 and West US 3. The errors look like permission problems, and the affected resources are becoming unavailable more frequently. Because this is happening across unrelated tenants, subscriptions, and resource groups, it doesn't look like a single service principal or credential-rotation issue.
Neither the public Azure Status page nor the customer-facing Service Health view is reporting an incident. We already have two Severity A support cases, but their response SLAs have been missed. Microsoft is now investigating, yet the lack of any status-page update has made it difficult to coordinate overnight coverage and plan for the production impact.
Has anyone dealt with a similar cross-tenant Storage outage? Also, how should Azure Status and Service Health be interpreted when a broad issue is affecting customers but neither page shows an incident?
3 Answers
The regional scope and simultaneous failures across unrelated tenants are important clues. Keep checking whether the errors are limited to a particular Storage account type, replication setting, API operation, or authentication path, and record timestamps in UTC. That information can help Support correlate the failures even while the status pages remain green.
The public Azure Status page isn’t a real-time dashboard for every problem. It generally shows incidents only after Microsoft determines they meet specific broad-impact criteria. A green indicator means there’s no published incident, not that every service is operating normally. Service Health is more targeted to the subscriptions and regions a customer uses, but it can still lag behind the first reports or omit issues that haven’t been formally identified yet.
This sounds more like a backend Storage issue than separate permission failures. Collect the exact Storage error codes and the x-ms-request-id or correlation ID from multiple affected tenants and regions, then attach those examples to the support cases. Showing the same failure pattern across unrelated customers is much stronger evidence than reporting isolated permission errors.
The customers have premium enterprise support, but we don’t directly own their support plans. We’ve been adding examples from each tenant to the existing Severity A cases.