Google Associate Cloud Engineer Practice Test 62
Question 610
An audit team needs to preserve a specific day's log entries before a log bucket retention cutoff. Which export action is appropriate?
- Create an appropriately protected log bucket with the required supported retention setting and route relevant logs there.
- Export the required entries to an approved durable destination using a scoped query and validate completeness and access.
- Disable audit-log collection to prevent the errors from being recorded.
- Delete the affected log bucket and assume missing historical entries can be recovered later.
Explanation
Correct Answer
B:Cloud Logging retention expiration is separate from long-term export; searches after expiry cannot reconstruct missing log entries.
Incorrect Answers
A:It does not provide the required log routing, evidence or retention setting. It leaves retention or recoverability controls unaddressed.
C:Disabling audit records suppresses evidence, not the underlying problem. The specific requirement concerns retention or recoverability controls.
D:It does not provide the required log routing, evidence or retention setting. The necessary distinction is retention or recoverability controls. The appropriate action here is: Export the required entries to an approved durable destination using a scoped query and validate completeness and access.
Question 611
A Cloud Logging sink was redirected to a newly created Cloud Storage bucket. Logs still enter Logs Explorer, but no objects appear in the new destination after its permissions were hardened. What should be checked?
- Grant the human operator Project Owner; the human user is always the identity exporting every sink event.
- Change the bucket storage class to Archive; the storage class automatically grants the Logging service account writer access.
- Grant the sink's actual writer identity the required permissions on the new destination and test routing with a fresh log entry.
- Reduce the number of log exclusions in the original project; exclusion settings repair any destination permission denial without IAM changes.
Explanation
Correct Answer
C:Grant the sink's actual writer identity the required permissions on the new destination and test routing with a fresh log entry. Export destinations must authorize the sink writer. Re-checking this precise binding and a fresh test log targets the failure introduced by destination hardening.
Incorrect Answers
A:Grant the human operator Project Owner; the human user is always the identity exporting every sink event. Logging sinks use a configured writer identity, often a service account. Human project roles do not necessarily authorize that sink to write the new destination.
B:Change the bucket storage class to Archive; the storage class automatically grants the Logging service account writer access. Storage class affects cost and access characteristics; it does not grant the sink principal permission to create output objects.
D:Reduce the number of log exclusions in the original project; exclusion settings repair any destination permission denial without IAM changes. Sink filters and exclusions can affect which entries are routed, but cannot override a rejected write from an unauthorized sink identity.
Question 612
An engineer uses an ingestion-time filter to exclude audit events and later tries to recover those excluded events from the same bucket. What limitation applies?
- Adjust the Monitoring filter to include intended live revisions or group safely by service/revision while preserving historical context.
- Disable the alerting policy so the current errors no longer appear.
- Restore the required ingestion path immediately and obtain any missing events from independent retained sources if available.
- Increase Cloud Logging retention as a replacement for the required alert or metric definition.
Explanation
Correct Answer
C:Entries removed by an exclusion filter may never have been stored in the target bucket and cannot be recovered by a new query.
Incorrect Answers
A:Monitoring reports symptoms; it does not enact the missing resource configuration. The necessary distinction is the relevant monitoring signal or alert.
B:Removing an alert conceals the failing signal instead of resolving its cause. The case instead depends on the relevant monitoring signal or alert.
D:Cloud Logging records telemetry, rather than directly enforcing the requested runtime behavior. It leaves the relevant monitoring signal or alert unaddressed. The appropriate action here is: Restore the required ingestion path immediately and obtain any missing events from independent retained sources if available.
Question 613
Cloud Audit Logs identify the caller but a developer assumes they record every byte returned by a BigQuery query. What is a defensible interpretation?
- Inspect the service-specific audit logging defaults and effective project or organization policy before changing audit configuration.
- Increase Cloud Logging retention as a replacement for the required alert or metric definition.
- Inspect a single average value and ignore distribution, errors and affected cohorts.
- Use audit metadata to trace access and use appropriate application or query-result controls for content-level evidence.
Explanation
Correct Answer
D:Audit logs document operations and principals but are not necessarily a full transcript of returned sensitive data.
Incorrect Answers
A:Changing audit collection cannot grant a denied permission or fix service behavior. The case instead depends on the correct audit-log evidence.
B:Cloud Logging records telemetry, rather than directly enforcing the requested runtime behavior. It leaves the correct audit-log evidence unaddressed.
C:It does not measure or notify on the signal requested by the case. The specific requirement concerns the correct audit-log evidence.
Question 614
One microservice emits timestamps without a timezone, causing unrelated logs to appear out of order in an incident timeline. What should be standardized?
- Reconstruct start/end times from provider advisories, timestamped logs and service metrics rather than repeating unsupported summary text.
- Disable the alerting policy so the current errors no longer appear.
- Rename the monitored VM and assume metric-series selection is automatically corrected.
- Emit UTC timestamps or unambiguous timezone-aware structured values and correlate events using request trace identifiers.
Explanation
Correct Answer
D:Unqualified timestamps from different hosts cannot be reliably ordered across time zones or skew without normalization.
Incorrect Answers
A:It does not measure or notify on the signal requested by the case. It leaves the relevant monitoring signal or alert unaddressed. The appropriate action here is: Emit UTC timestamps or unambiguous timezone-aware structured values and correlate events using request trace identifiers.
B:Removing an alert conceals the failing signal instead of resolving its cause. The specific requirement concerns the relevant monitoring signal or alert.
C:Renaming a resource does not alter the actual effective configuration. The necessary distinction is the relevant monitoring signal or alert.
Question 615
A trace graph shows a database span consuming most request time, but its service owner only optimizes the front-end container CPU. What targeted investigation should be done?
- Trace the end-to-end request through gateway and backend spans using correlated request identifiers.
- Inspect a single average value and ignore distribution, errors and affected cohorts.
- Inspect database span attributes, query timing and related slow-query evidence before changing the front end.
- Increase Cloud Logging retention as a replacement for the required alert or metric definition.
Explanation
Correct Answer
C:Distributed traces identify which downstream operation dominates request latency; unrelated front-end CPU tuning is unlikely to fix a database bottleneck.
Incorrect Answers
A:It does not measure or notify on the signal requested by the case. The specific requirement concerns the relevant monitoring signal or alert. The appropriate action here is: Inspect database span attributes, query timing and related slow-query evidence before changing the front end.
B:It does not measure or notify on the signal requested by the case. The necessary distinction is the relevant monitoring signal or alert. The appropriate action here is: Inspect database span attributes, query timing and related slow-query evidence before changing the front end.
D:Cloud Logging records telemetry, rather than directly enforcing the requested runtime behavior. The case instead depends on the relevant monitoring signal or alert. The appropriate action here is: Inspect database span attributes, query timing and related slow-query evidence before changing the front end.
Question 616
A request travels from Cloud Run through an API gateway to a GKE backend. Traces show separate root spans because the gateway strips the incoming trace context. What must be preserved for one end-to-end trace?
- Propagate the supported distributed-tracing trace context headers through the gateway and verify that backend spans share the incoming trace ID.
- Increase Cloud Run request timeout; timeout values automatically link spans that were started without a shared context.
- Use the same Cloud Billing account for both services; trace IDs are generated from billing-account association.
- Rename the gateway hostname; identical service hostnames automatically merge unrelated trace roots after ingestion.
Explanation
Correct Answer
A:Propagate the supported distributed-tracing trace context headers through the gateway and verify that backend spans share the incoming trace ID. Distributed trace context carries the trace identity and parent relationship across service boundaries. A gateway that preserves it allows correlated spans.
Incorrect Answers
B:Increase Cloud Run request timeout; timeout values automatically link spans that were started without a shared context. Request duration settings do not convey a parent trace ID from one process to another and cannot repair stripped propagation headers.
C:Use the same Cloud Billing account for both services; trace IDs are generated from billing-account association. Billing association is independent from the trace context used to correlate RPCs across Cloud Run and Kubernetes services.
D:Rename the gateway hostname; identical service hostnames automatically merge unrelated trace roots after ingestion. Hostname names are not a substitute for transmitting trace IDs between individual calls. Spans stay separate without propagation.
Question 617
An on-call engineer samples only successful low-latency requests, and distributed traces omit rare 30-second errors. What instrumentation policy needs review?
- Use an appropriate sampling strategy that preserves a representative set of failures and high-latency traces within privacy budgets.
- Monitor tail-latency percentiles such as P95 or P99 and error rates for representative request cohorts.
- Rename the monitored VM and assume metric-series selection is automatically corrected.
- Set an unrelated CPU threshold without selecting the actual signal or metric labels.
Explanation
Correct Answer
A:Trace sampling can systematically hide tail incidents when it excludes errors or slow requests.
Incorrect Answers
B:It does not measure or notify on the signal requested by the case. The case instead depends on the requested latency measurement.
C:Renaming a resource does not alter the actual effective configuration. It leaves the requested latency measurement unaddressed.
D:It does not measure or notify on the signal requested by the case. The specific requirement concerns the requested latency measurement. The appropriate action here is: Use an appropriate sampling strategy that preserves a representative set of failures and high-latency traces within privacy budgets.
Question 618
A development team adds an attribute containing raw customer names to every trace span and violates data-minimization policy. Which remediation is necessary?
- Inspect database span attributes, query timing and related slow-query evidence before changing the front end.
- Disable the alerting policy so the current errors no longer appear.
- Set an unrelated CPU threshold without selecting the actual signal or metric labels.
- Remove or tokenize personally identifiable span attributes and restrict collection to approved diagnostics.
Explanation
Correct Answer
D:Trace attributes are telemetry data subject to privacy controls; convenience does not justify collecting customer identifiers.
Incorrect Answers
A:It does not measure or notify on the signal requested by the case. It leaves the relevant monitoring signal or alert unaddressed. The appropriate action here is: Remove or tokenize personally identifiable span attributes and restrict collection to approved diagnostics.
B:Removing an alert conceals the failing signal instead of resolving its cause. The specific requirement concerns the relevant monitoring signal or alert. The appropriate action here is: Remove or tokenize personally identifiable span attributes and restrict collection to approved diagnostics.
C:It does not measure or notify on the signal requested by the case. The necessary distinction is the relevant monitoring signal or alert. The appropriate action here is: Remove or tokenize personally identifiable span attributes and restrict collection to approved diagnostics.
Question 619
A service's trace samples point to CPU time concentrated in a serialization library, but the team plans to increase an unrelated SQL database tier. What tool offers code-level evidence?
- Inspect database span attributes, query timing and related slow-query evidence before changing the front end.
- Rename the monitored VM and assume metric-series selection is automatically corrected.
- Use supported CPU profiling or application instrumentation to identify the hot serialization path before a code optimization.
- Inspect a single average value and ignore distribution, errors and affected cohorts.
Explanation
Correct Answer
C:A performance profiler can attribute CPU consumption to functions and call stacks, unlike a database-capacity dashboard.
Incorrect Answers
A:It does not measure or notify on the signal requested by the case. The specific requirement concerns the relevant monitoring signal or alert. The appropriate action here is: Use supported CPU profiling or application instrumentation to identify the hot serialization path before a code optimization.
B:Renaming a resource does not alter the actual effective configuration. The necessary distinction is the relevant monitoring signal or alert. The appropriate action here is: Use supported CPU profiling or application instrumentation to identify the hot serialization path before a code optimization.
D:It does not measure or notify on the signal requested by the case. The case instead depends on the relevant monitoring signal or alert. The appropriate action here is: Use supported CPU profiling or application instrumentation to identify the hot serialization path before a code optimization.
Question 620
A Java process reports memory growth after every request but aggregated Cloud Monitoring memory graphs cannot identify the allocation source. What additional evidence should be collected?
- Use one documented metric unit and perform explicit unit conversion at the exporters; verify monitored-resource and metric labels.
- Rename the monitored VM and assume metric-series selection is automatically corrected.
- Collect a supported application memory profile or heap diagnostics and correlate them with traffic and releases.
- Send every alert to all organization members instead of configuring intended notification channels.
Explanation
Correct Answer
C:Host memory utilization describes resource usage but not which application allocation path leaks memory.
Incorrect Answers
A:It does not measure or notify on the signal requested by the case. The necessary distinction is the relevant monitoring signal or alert. The appropriate action here is: Collect a supported application memory profile or heap diagnostics and correlate them with traffic and releases.
B:Renaming a resource does not alter the actual effective configuration. The case instead depends on the relevant monitoring signal or alert. The appropriate action here is: Collect a supported application memory profile or heap diagnostics and correlate them with traffic and releases.
D:It does not measure or notify on the signal requested by the case. It leaves the relevant monitoring signal or alert unaddressed. The appropriate action here is: Collect a supported application memory profile or heap diagnostics and correlate them with traffic and releases.