Question 70
A development group must find why agents fail to complete workflows despite healthy model endpoints. What is the most appropriate Azure AI implementation?
- Track tool invocation errors and timeouts
- Trace each retrieval and generation span
- Monitor token utilization across agent turns
- Compare behavior across model or prompt versions
Explanation
Correct Answer
A:For the stated requirement to find why agents fail to complete workflows despite healthy model endpoints, tool telemetry records HTTP failures, permission denials and timeouts at action boundaries; model token metrics will not diagnose a downstream API error; this is the mechanism that the system must implement.
Incorrect Answers
B:For a system whose objective is to locate whether a slow RAG response is delayed by search, tool calls, or model generation, distributed traces expose timing and relationships among retrieval, tool and inference spans; aggregate CPU graphs alone cannot isolate these steps; this does not directly meet the required outcome to find why agents fail to complete workflows despite healthy model endpoints.
C:For a system whose objective is to identify waste when conversation history and tool outputs grow over successive exchanges, per-turn token accounting reveals input-context inflation; controlled summarization or trimming can reduce cost while preserving vital state; this does not directly meet the required outcome to find why agents fail to complete workflows despite healthy model endpoints.
D:For a system whose objective is to detect regression when a newly deployed model changes how it answers the same business questions, version-tagged test results separate model changes from request-mix changes and allow a reliable rollback decision; this does not directly meet the required outcome to find why agents fail to complete workflows despite healthy model endpoints.
Question 71
A solution architect is deciding how to detect regression when a newly deployed model changes how it answers the same business questions. Which Azure AI pattern addresses that task?
- Inspect embedding pipeline health after model changes
- Limit prompt and retrieved context length
- Monitor token utilization across agent turns
- Separate data-plane use from resource management
- Compare behavior across model or prompt versions
- Monitor retrieval freshness and ingestion errors
Explanation
Correct Answer
E:For the stated requirement to detect regression when a newly deployed model changes how it answers the same business questions, version-tagged test results separate model changes from request-mix changes and allow a reliable rollback decision; this is the mechanism that the system must implement.
Incorrect Answers
A:For a system whose objective is to catch incompatible document and query vectors when retrieval quality suddenly collapses, embedding dimensionality and model compatibility must be maintained between indexed passages and query vectors; vector space changes require a coordinated migration; this does not directly meet the required outcome to detect regression when a newly deployed model changes how it answers the same business questions.
B:For a system whose objective is to reduce inference spend caused by oversized repeated document context, input tokens contribute to context limits and usage costs; pruning irrelevant evidence can improve efficiency without changing the model deployment; this does not directly meet the required outcome to detect regression when a newly deployed model changes how it answers the same business questions.
C:For a system whose objective is to identify waste when conversation history and tool outputs grow over successive exchanges, per-turn token accounting reveals input-context inflation; controlled summarization or trimming can reduce cost while preserving vital state; this does not directly meet the required outcome to detect regression when a newly deployed model changes how it answers the same business questions.
D:For a system whose objective is to let operators change infrastructure while limiting which identities can read sensitive index documents, management-plane roles and data-plane roles authorize different operations; control over deployment configuration does not imply every content read permission; this does not directly meet the required outcome to detect regression when a newly deployed model changes how it answers the same business questions.
F:For a system whose objective is to detect when newly approved source documents are not available to the grounding pipeline, indexer execution status, document counts, and source timestamps show whether retrieval data is current; a good model response cannot compensate for missing content; this does not directly meet the required outcome to detect regression when a newly deployed model changes how it answers the same business questions.
Question 72
The planned deployment includes a mechanism to monitor token utilization across agent turns. Which outcome should the QA team measure?
- Compare indexed versions and document counts with the authoritative document repository
- Validate vector dimensions and semantic recall using a fixed representative query set
- Plot tool error rates and p95 tool latency against end-to-end task completion
- Graph cost per completed task and token usage across dialogue length bands
Explanation
Correct Answer
D:The accepted change is to monitor token utilization across agent turns; this evaluation checks whether it can identify waste when conversation history and tool outputs grow over successive exchanges, since per-turn token accounting reveals input-context inflation; controlled summarization or trimming can reduce cost while preserving vital state.
Incorrect Answers
A:This test concerns the ability to detect when newly approved source documents are not available to the grounding pipeline, for which indexer execution status, document counts, and source timestamps show whether retrieval data is current; a good model response cannot compensate for missing content; it does not establish whether the change to monitor token utilization across agent turns works for the current release.
B:This test concerns the ability to catch incompatible document and query vectors when retrieval quality suddenly collapses, for which embedding dimensionality and model compatibility must be maintained between indexed passages and query vectors; vector space changes require a coordinated migration; it does not establish whether the change to monitor token utilization across agent turns works for the current release.
C:This test concerns the ability to find why agents fail to complete workflows despite healthy model endpoints, for which tool telemetry records HTTP failures, permission denials and timeouts at action boundaries; model token metrics will not diagnose a downstream API error; it does not establish whether the change to monitor token utilization across agent turns works for the current release.
Question 73
A user-acceptance test fails with the following observation: Concurrent clients immediately retry every throttled request, increasing the rate of failures. What is the best correction?
- Cache validated FAQ outputs by normalized question and content version with proper expiration
- Remove generic outbound reachability and allow only approved tool destinations and methods
- Implement bounded retry-after-aware backoff with jitter and limit client concurrency
- Trim and rerank context, bound retrieval count, and retain only task-relevant evidence
Explanation
Correct Answer
C:The fault signal is that concurrent clients immediately retry every throttled request, increasing the rate of failures; the targeted fix works because exponential backoff with jitter and respect for retry hints reduces synchronized retries; extra retries cannot replace insufficient sustained capacity, restoring the ability to recover from temporary HTTP 429 responses without amplifying load on the model deployment.
Incorrect Answers
A:If the failing function were to avoid repeating the same approved reference answer for unchanged public content, this change could be relevant because an application cache can reduce repeated inference requests when answers and policy versions are safe to reuse; invalidation and access scope remain essential; the actual symptom is that concurrent clients immediately retry every throttled request, increasing the rate of failures, which calls for a different repair.
B:If the failing function were to prevent a compromised agent from calling unapproved external endpoints, this change could be relevant because tool allowlists, network egress policies, and scoped identities limit what an agent can reach; prompt instructions alone are not a network control; the actual symptom is that concurrent clients immediately retry every throttled request, increasing the rate of failures, which calls for a different repair.
D:If the failing function were to reduce inference spend caused by oversized repeated document context, this change could be relevant because input tokens contribute to context limits and usage costs; pruning irrelevant evidence can improve efficiency without changing the model deployment; the actual symptom is that concurrent clients immediately retry every throttled request, increasing the rate of failures, which calls for a different repair.
Question 74
A release reviewer is assessing a feature intended to trace each retrieval and generation span. Which test is most meaningful?
- Confirm no requests use the retired credential and access continues during rotation
- Compare indexed versions and document counts with the authoritative document repository
- Track policy category, severity, false-positive review, and block rate by version
- Inspect per-stage p95 latency and correlation identifiers for slow requests
Explanation
Correct Answer
D:The accepted change is to trace each retrieval and generation span; this evaluation checks whether it can locate whether a slow RAG response is delayed by search, tool calls, or model generation, since distributed traces expose timing and relationships among retrieval, tool and inference spans; aggregate CPU graphs alone cannot isolate these steps.
Incorrect Answers
A:This test concerns the ability to replace an exposed secret while keeping a production inference endpoint available, for which credential rollover with overlapping validity and centralized secret distribution enables staged updates; longer-term managed identity can reduce secret use; it does not establish whether the change to trace each retrieval and generation span works for the current release.
B:This test concerns the ability to detect when newly approved source documents are not available to the grounding pipeline, for which indexer execution status, document counts, and source timestamps show whether retrieval data is current; a good model response cannot compensate for missing content; it does not establish whether the change to trace each retrieval and generation span works for the current release.
C:This test concerns the ability to identify spikes in unsafe or blocked prompts and unsafe generated material, for which guardrail and content-filter telemetry reveals policy-related events; it should be segmented without exposing sensitive payloads unnecessarily; it does not establish whether the change to trace each retrieval and generation span works for the current release.
Question 75
A pilot has a functional gap: A leaked static key appears in a build log, but immediate invalidation would break running replicas. How should the team address that gap?
- Remove generic outbound reachability and allow only approved tool destinations and methods
- Grant the specific Search index data-reading permission at the appropriate resource scope
- Trim and rerank context, bound retrieval count, and retain only task-relevant evidence
- Introduce a new credential through a secret store, roll replicas, then revoke the exposed credential
Explanation
Correct Answer
D:The fault signal is that a leaked static key appears in a build log, but immediate invalidation would break running replicas; the targeted fix works because credential rollover with overlapping validity and centralized secret distribution enables staged updates; longer-term managed identity can reduce secret use, restoring the ability to replace an exposed secret while keeping a production inference endpoint available.
Incorrect Answers
A:If the failing function were to prevent a compromised agent from calling unapproved external endpoints, this change could be relevant because tool allowlists, network egress policies, and scoped identities limit what an agent can reach; prompt instructions alone are not a network control; the actual symptom is that a leaked static key appears in a build log, but immediate invalidation would break running replicas, which calls for a different repair.
B:If the failing function were to allow a retrieval application to query only the approved search resource without subscription administration rights, this change could be relevant because Azure RBAC can grant scoped data-plane permissions to the workload principal; broad management roles are unnecessary for query-only service calls; the actual symptom is that a leaked static key appears in a build log, but immediate invalidation would break running replicas, which calls for a different repair.
C:If the failing function were to reduce inference spend caused by oversized repeated document context, this change could be relevant because input tokens contribute to context limits and usage costs; pruning irrelevant evidence can improve efficiency without changing the model deployment; the actual symptom is that a leaked static key appears in a build log, but immediate invalidation would break running replicas, which calls for a different repair.
Question 76
An engineering team has a requirement to allow a retrieval application to query only the approved search resource without subscription administration rights. Which capability should be in the initial design?
- Restrict agent outbound tool access
- Rotate compromised credentials without app downtime
- Inspect embedding pipeline health after model changes
- Assign least-privilege data access to a workload identity
Explanation
Correct Answer
D:For the stated requirement to allow a retrieval application to query only the approved search resource without subscription administration rights, Azure RBAC can grant scoped data-plane permissions to the workload principal; broad management roles are unnecessary for query-only service calls; this is the mechanism that the system must implement.
Incorrect Answers
A:For a system whose objective is to prevent a compromised agent from calling unapproved external endpoints, tool allowlists, network egress policies, and scoped identities limit what an agent can reach; prompt instructions alone are not a network control; this does not directly meet the required outcome to allow a retrieval application to query only the approved search resource without subscription administration rights.
B:For a system whose objective is to replace an exposed secret while keeping a production inference endpoint available, credential rollover with overlapping validity and centralized secret distribution enables staged updates; longer-term managed identity can reduce secret use; this does not directly meet the required outcome to allow a retrieval application to query only the approved search resource without subscription administration rights.
C:For a system whose objective is to catch incompatible document and query vectors when retrieval quality suddenly collapses, embedding dimensionality and model compatibility must be maintained between indexed passages and query vectors; vector space changes require a coordinated migration; this does not directly meet the required outcome to allow a retrieval application to query only the approved search resource without subscription administration rights.
Question 77
Stakeholders want a production AI solution able to serve a stable higher request rate when the deployment repeatedly reaches its assigned token budget. Which choice implements that requirement?
- Inspect embedding pipeline health after model changes
- Monitor safety-related request outcomes
- Adjust quota allocation for sustained traffic
- Use caching for reusable deterministic outputs
- Limit prompt and retrieved context length
Explanation
Correct Answer
C:For the stated requirement to serve a stable higher request rate when the deployment repeatedly reaches its assigned token budget, tokens-per-minute and request limits constrain throughput; allocation or capacity decisions must match the deployment's actual measured demand; this is the mechanism that the system must implement.
Incorrect Answers
A:For a system whose objective is to catch incompatible document and query vectors when retrieval quality suddenly collapses, embedding dimensionality and model compatibility must be maintained between indexed passages and query vectors; vector space changes require a coordinated migration; this does not directly meet the required outcome to serve a stable higher request rate when the deployment repeatedly reaches its assigned token budget.
B:For a system whose objective is to identify spikes in unsafe or blocked prompts and unsafe generated material, guardrail and content-filter telemetry reveals policy-related events; it should be segmented without exposing sensitive payloads unnecessarily; this does not directly meet the required outcome to serve a stable higher request rate when the deployment repeatedly reaches its assigned token budget.
D:For a system whose objective is to avoid repeating the same approved reference answer for unchanged public content, an application cache can reduce repeated inference requests when answers and policy versions are safe to reuse; invalidation and access scope remain essential; this does not directly meet the required outcome to serve a stable higher request rate when the deployment repeatedly reaches its assigned token budget.
E:For a system whose objective is to reduce inference spend caused by oversized repeated document context, input tokens contribute to context limits and usage costs; pruning irrelevant evidence can improve efficiency without changing the model deployment; this does not directly meet the required outcome to serve a stable higher request rate when the deployment repeatedly reaches its assigned token budget.
Question 78
Production telemetry reveals this failure: A new query embedder is deployed without reindexing and semantically relevant matches disappear. Which change most directly corrects it?
- Tag outputs by model and prompt version and run side-by-side regression cases
- Instrument each tool call with duration, status, and bounded retry decisions
- Align query embeddings with index embeddings and re-embed stored content when migrating models
- Run source-grounded evaluations and inspect failing claim-to-passage alignments
Explanation
Correct Answer
C:The fault signal is that a new query embedder is deployed without reindexing and semantically relevant matches disappear; the targeted fix works because embedding dimensionality and model compatibility must be maintained between indexed passages and query vectors; vector space changes require a coordinated migration, restoring the ability to catch incompatible document and query vectors when retrieval quality suddenly collapses.
Incorrect Answers
A:If the failing function were to detect regression when a newly deployed model changes how it answers the same business questions, this change could be relevant because version-tagged test results separate model changes from request-mix changes and allow a reliable rollback decision; the actual symptom is that a new query embedder is deployed without reindexing and semantically relevant matches disappear, which calls for a different repair.
B:If the failing function were to find why agents fail to complete workflows despite healthy model endpoints, this change could be relevant because tool telemetry records HTTP failures, permission denials and timeouts at action boundaries; model token metrics will not diagnose a downstream API error; the actual symptom is that a new query embedder is deployed without reindexing and semantically relevant matches disappear, which calls for a different repair.
D:If the failing function were to detect fluent answers that introduce claims unsupported by retrieved passages, this change could be relevant because groundedness evaluations compare generated claims with supplied evidence; fluency or user satisfaction alone cannot establish factual support; the actual symptom is that a new query embedder is deployed without reindexing and semantically relevant matches disappear, which calls for a different repair.
Question 79
Production telemetry reveals this failure: A popular public FAQ triggers identical generation repeatedly despite no source changes. Which change most directly corrects it?
- Alert on ingestion failures and index freshness drift, then repair the source pipeline
- Cache validated FAQ outputs by normalized question and content version with proper expiration
- Implement bounded retry-after-aware backoff with jitter and limit client concurrency
- Review quotas and reallocate or request capacity for the applicable model and region
Explanation
Correct Answer
B:The fault signal is that a popular public FAQ triggers identical generation repeatedly despite no source changes; the targeted fix works because an application cache can reduce repeated inference requests when answers and policy versions are safe to reuse; invalidation and access scope remain essential, restoring the ability to avoid repeating the same approved reference answer for unchanged public content.
Incorrect Answers
A:If the failing function were to detect when newly approved source documents are not available to the grounding pipeline, this change could be relevant because indexer execution status, document counts, and source timestamps show whether retrieval data is current; a good model response cannot compensate for missing content; the actual symptom is that a popular public FAQ triggers identical generation repeatedly despite no source changes, which calls for a different repair.
C:If the failing function were to recover from temporary HTTP 429 responses without amplifying load on the model deployment, this change could be relevant because exponential backoff with jitter and respect for retry hints reduces synchronized retries; extra retries cannot replace insufficient sustained capacity; the actual symptom is that a popular public FAQ triggers identical generation repeatedly despite no source changes, which calls for a different repair.
D:If the failing function were to serve a stable higher request rate when the deployment repeatedly reaches its assigned token budget, this change could be relevant because tokens-per-minute and request limits constrain throughput; allocation or capacity decisions must match the deployment's actual measured demand; the actual symptom is that a popular public FAQ triggers identical generation repeatedly despite no source changes, which calls for a different repair.
Question 80
A new production AI solution must detect fluent answers that introduce claims unsupported by retrieved passages. Which approach should the team select first?
- Trace each retrieval and generation span
- Monitor safety-related request outcomes
- Measure groundedness with cited source evidence
- Restrict agent outbound tool access
Explanation
Correct Answer
C:For the stated requirement to detect fluent answers that introduce claims unsupported by retrieved passages, groundedness evaluations compare generated claims with supplied evidence; fluency or user satisfaction alone cannot establish factual support; this is the mechanism that the system must implement.
Incorrect Answers
A:For a system whose objective is to locate whether a slow RAG response is delayed by search, tool calls, or model generation, distributed traces expose timing and relationships among retrieval, tool and inference spans; aggregate CPU graphs alone cannot isolate these steps; this does not directly meet the required outcome to detect fluent answers that introduce claims unsupported by retrieved passages.
B:For a system whose objective is to identify spikes in unsafe or blocked prompts and unsafe generated material, guardrail and content-filter telemetry reveals policy-related events; it should be segmented without exposing sensitive payloads unnecessarily; this does not directly meet the required outcome to detect fluent answers that introduce claims unsupported by retrieved passages.
D:For a system whose objective is to prevent a compromised agent from calling unapproved external endpoints, tool allowlists, network egress policies, and scoped identities limit what an agent can reach; prompt instructions alone are not a network control; this does not directly meet the required outcome to detect fluent answers that introduce claims unsupported by retrieved passages.