Azure AI Cloud Developer Associate (AI-200) - Full Practice Exam 1

1. A backend creates many short-lived PostgreSQL connections and encounters connection pressure. Which capability should it evaluate?
2. A backend processes business commands asynchronously and needs message locks, dead-letter handling, and supported ordering for related commands. Which service should it evaluate?
3. A background consumer logs an error but the originating request cannot be identified. Which improvement is most relevant?
4. A cached account balance becomes obsolete after a successful update. What should the application do?
5. A change-feed handler submits an external action and can receive the same change again after recovery. Which property should the action handling provide?
6. A Container Apps release must expose a new version to a controlled fraction of HTTP traffic while the old revision remains available. Which capability fits?
7. A container process starts but cannot yet serve requests because initialization is still running. Which probe should control whether it receives traffic?
8. A Cosmos-backed AI application stores embeddings and performs similarity retrieval. Which relationship must be maintained?
9. A Cosmos query filters a property that is not appropriately indexed and has high RU cost. What should be evaluated?
10. A downstream service credential is rotated in Key Vault, but application instances keep using the old cached value indefinitely. What should be improved?
11. A feature must be enabled for a controlled subset of users without rebuilding the container image. Which approach fits?
12. A Function should execute when a supported queue receives a message and write to another supported service with minimal plumbing. Which concepts should be configured?
13. A Function's trigger is configured correctly, but access to its target service fails. What must still be validated independently?
14. A handler may legitimately run longer than the current message lock duration. What should the consumer design evaluate?
15. A Kubernetes workload's Pod addresses change during updates. Internal consumers need a stable logical endpoint for the application. Which resource normally provides it?
16. A model-backed API requires an alert when sustained latency breaches its operating threshold. Which configuration should be used?
17. A pgvector query must retrieve the nearest documents from one authorized tenant. Which design is appropriate?
18. A poison message repeatedly fails because required business fields are invalid. What should the consumer do rather than retry it forever?
19. A PostgreSQL application needs vector similarity queries using the supported pgvector extension. What should the deployment verify first?
20. A production container release must reproduce the tested artifact even if a tag later moves. Which TWO practices fit? Select TWO.
21. A queue-triggered Function performs a consequential operation. The team expects retries after transient failures. Which implementation property is essential?
22. A Redis key must expire after a specified lifetime. Which feature should the application use?
23. A request crosses an API, queue consumer, and database. Which TWO telemetry practices support end-to-end latency attribution? Select TWO.
24. A service call fails because a required parameter is invalid. Which retry decision is appropriate?
25. A service reacts to new or updated Cosmos items. It scales the processor across workers and must resume after failures. What should be configured?
26. A service should pick up approved configuration changes while running. What must its integration implement?
27. A shared semantic cache serves multiple tenants. Which requirement is essential before reusing a cached result?
28. A team needs centrally managed nonsecret configuration and feature flags across application instances. Which service should it evaluate?
29. A team needs to build and retain private OCI images for Azure-hosted applications. Which service should it evaluate?
30. A team wants supported cloud-based container builds triggered by source or base-image changes without maintaining its own build VM. Which capability fits?
31. A worker completes an external operation but crashes before settling the Service Bus message. The message is delivered again. Which design avoids duplicate business effects?
32. AKS Pods fail with image-pull authorization errors against a private registry. What should be checked first?
33. An AI API repeatedly retrieves unchanged reference data and can tolerate a defined freshness window. Which pattern can reduce database load?
34. An AI result cache serves obsolete policy answers after the underlying documents change. Which design should be improved?
35. An AI service knows a stored conversation item's id and partition key. Which operation avoids unnecessary query fan-out?
36. An App Service container needs different configuration in test and production without rebuilding its image. What should the deployment supply?
37. An application must run multiple identical Kubernetes replicas and roll out new versions declaratively. Which Kubernetes resource should be used for the stateless application controller?
38. An application needs event-driven reactions to storage changes with filtered subscriptions. Which service is designed for this routing pattern?
39. An application wants vector similarity retrieval in an eligible Redis deployment. What should it configure and verify?
40. An Azure application needs a secret from Key Vault without a hardcoded client secret. Which design should it use where supported?
41. An engineer needs to group recent error records by service and count occurrences in Azure Monitor Logs. Which language should be used?
42. An Event Grid handler may receive repeated deliveries. What should the application's handler assume?
43. An HTTP container app scales to zero and the first request after idle periods is slower. Which design option should be evaluated if the latency requirement cannot tolerate cold starts?
44. Approximate vector search is faster but misses some relevant neighbors. Which tradeoff should the developer evaluate?
45. Background containers should scale according to the backlog in a supported queue. Which Container Apps scaling capability should be configured?
46. Commands for the same order need ordered handling while independent orders may process concurrently. Which Service Bus feature should be considered?
47. Diagnostic logging currently records prompts containing sensitive customer data and raw credentials. What should the team change?
48. Pods repeatedly restart after a deployment. Which evidence is most useful for initial diagnosis?
49. Several independent services must receive their own eligible copy of published business messages. Which Service Bus pattern fits?
50. Vector queries compete with transactional work for CPU and memory. What should be measured before choosing a new PostgreSQL compute configuration?

Comments — 0