Free Microsoft Certified: Azure Developer Associate practice — 6 questions on Develop Azure compute solutions, with explanations. No sign-up.
Full 12-question mixed test →
Question 1 of 6 · Develop Azure compute solutions
An orchestrator function in a Durable Functions app must wait for an external HTTP callback approving a purchase order. If no approval arrives within 24 hours, the orchestration must automatically escalate to a manager. Which implementation correctly achieves this?
The correct pattern is to create a task from ctx.WaitForExternalEvent for the approval event and a separate durable timer task (ctx.CreateTimer) for the 24-hour deadline, then use Task.WhenAny to see which completes first. If the timer wins, the orchestrator cancels the wait and escalates. This is the documented 'Human interaction / timeout' pattern in Durable Functions.
Question 2 of 6 · Develop Azure compute solutions
A Container App consumes messages from an Azure Service Bus queue. Requirements: scale to zero replicas when the queue is empty to minimize cost, and scale out up to 10 replicas as queue length grows. Which scaling configuration meets these requirements?
Azure Container Apps uses KEDA scalers under the hood. The azure-servicebus scale rule type natively reads queue length and can scale replicas to zero when the queue is empty and up to maxReplicas as messages accumulate — this is the documented, purpose-built scaler for this exact scenario.
Question 3 of 6 · Develop Azure compute solutions
A web app has staging and production deployment slots. The connection string points to a test database in staging and the production database in production. After every deployment, staging is swapped into production, but the team wants the connection string to NEVER be swapped — each slot must always keep its own value regardless of how many swaps occur. What must be configured?
A setting marked as a deployment slot setting ('sticky') stays with its slot and is never moved during a swap. To guarantee both staging and production always retain their own distinct connection string values across repeated swaps, the sticky flag must be set on the setting in EACH slot, not just one.
Question 4 of 6 · Develop Azure compute solutions
A company is building a latency-sensitive Azure Function that must access a SQL Database deployed with a private endpoint inside a virtual network, and it cannot tolerate cold-start delays on the first invocation after idle periods. Which hosting plan should be used?
The Elastic Premium plan supports pre-warmed ('always ready') instances that eliminate cold starts, and it supports regional VNet integration to reach resources like a private-endpoint-protected SQL Database — satisfying both requirements simultaneously.
Question 5 of 6 · Develop Azure compute solutions
An App Service has autoscale rules configured to scale out when CPU exceeds 70% and scale in when CPU drops below 30%. In production, the instance count is oscillating rapidly — scaling out then immediately back in every few minutes — even though load is only moderately variable. What should be changed to fix this flapping behavior?
Flapping is a classic symptom of a cooldown period that's too short, causing autoscale to react to the same transient metric spike or normal fluctuation repeatedly. Increasing the cooldown period (the default is 5 minutes) between successive scale actions gives instances time to stabilize before the next evaluation, which directly resolves oscillation.
Question 6 of 6 · Develop Azure compute solutions
A team maintains a custom application container image built FROM a Microsoft-provided base OS image in Azure Container Registry. They need the application image to be automatically rebuilt and pushed to ACR whenever Microsoft releases a security patch to that base OS image — with zero manual steps. Which ACR Tasks configuration achieves this?
ACR Tasks support a base image update trigger, which detects when the base image referenced in a Dockerfile is updated (in ACR or a supported public registry) and automatically triggers a rebuild and push of the dependent image — with no manual intervention required.
Ready for the real thing?
The full course has two full-length practice tests, video lessons for every exam domain, hands-on labs and detailed answer explanations.