Microsoft's November 16, 2026 announcement introduces task queues in Durable Task Scheduler public preview. A queue groups work inside a task hub and can pace how quickly the scheduler hands it to workers. Microsoft announcement.
My proposed pilot would start with one downstream dependency and a written throughput objective. Record the workload, worker configuration, expected response behavior, and the owner who will evaluate the test.
Understand What the Rate Measures
Microsoft says the rate counts scheduler work-item handouts, including orchestration turns and activity executions. The queue rate is not the number of completed orchestrations; bursts and retries can occur. Queue behavior.
I would measure handouts, completed work, downstream requests, retries, and response times separately. Use synthetic work in an approved test environment and preserve the observation window with the results.
Review Routing and Workers
Activities and sub-orchestrations inherit their parent's queue unless overridden. Microsoft says work in a defined queue requires an unfiltered worker, and a missing queue falls back to normal routing with a warning. Routing and worker requirements.
My acceptance sheet would identify the queue name, routing decision, and worker configuration. Include a controlled missing-queue case and verify the resulting warning and workload behavior before relying on the proposed setup.
Protect a Hard Downstream Limit
The announcement says queue pacing does not provide an exact downstream request cap because bursts and retries can occur. For a hard limit, Microsoft recommends additional rate limiting in the calling activity. Preview limitations.
Have the downstream service owner define the applicable constraint and acceptance criteria. Compare observed request rates and error responses with that requirement, and investigate any mismatch before expanding the pilot.
Plan Lifecycle Operations
Microsoft notes that deleting a queue moves pending work to normal routing rather than waiting for it to drain. Extended sessions are not paced in this preview. Operational limits.
I would record who may change a queue, how pending work is observed, and the approved recovery procedure. Rehearse a configuration change with disposable work and confirm both scheduler state and the downstream application's outcome.
Practical Cloud Engineer Takeaway
Measure scheduler handouts and downstream requests independently.
Verify queue routing and unfiltered worker availability.
Retain an activity-level control for hard downstream limits.
Rehearse missing-queue and configuration-change behavior.
Document ownership and recovery before wider use.
Who Should Care?
Durable Task Scheduler developers, Azure Functions teams, and platform engineers managing workflows that call rate-limited dependencies.
Bottom Line
Task queues add a pacing option for scheduler work. A useful pilot demonstrates the actual downstream request pattern and makes the preview's routing and worker boundaries visible to operators.
No task hub, queue, worker, workflow, or Azure resource was changed for this article.
Sources
Stay radical, stay curious, and keep pushing the boundaries of what is possible in the cloud.
Chriz
Beyond Cloud with Chriz
Comments