Google Associate Cloud Engineer Practice Test 28
Question 270
A globally deployed financial service requires horizontally scalable relational transactions with strong consistency across multiple locations. Which database should be evaluated?
- Use Cloud SQL as an unreplicated single-region database for the global transaction requirement.
- Choose an appropriate Spanner instance configuration with required consistency and geographic placement.
- Use BigQuery scheduled queries in place of online relational transactions.
- Use Firestore documents instead of the required relational schema and SQL behavior.
Explanation
Correct Answer
B:Spanner combines relational transactions and horizontal scalability with strongly consistent replication.
Incorrect Answers
A:Placement changes do not automatically implement the required resource-level feature. It leaves the transactional schema or consistency requirement unaddressed.
C:It does not supply the specified transactional schema or consistency behavior. The specific requirement concerns the transactional schema or consistency requirement.
D:It does not supply the specified transactional schema or consistency behavior. The necessary distinction is the transactional schema or consistency requirement.
Question 271
An IoT platform inserts billions of timestamp-keyed sensor observations and requires low-latency reads by row key, not ad hoc joins. Which database fits?
- Use Bigtable with row-key design optimized for the expected read/write patterns.
- Use Spanner SQL queries as a drop-in replacement for the specified low-latency row-key access pattern.
- Store all individual observations in BigQuery and assume its analytical engine provides the required point-read latency.
- Use Cloud SQL with one monolithic transactional table for unlimited IoT writes without assessing scale.
Explanation
Correct Answer
A:Bigtable is a managed wide-column store designed for high-throughput, low-latency key-based access at scale.
Incorrect Answers
B:It does not correct the Bigtable row-key, cluster or data-access requirement. The specific requirement concerns Bigtable row-key distribution.
C:It does not correct the Bigtable row-key, cluster or data-access requirement. The necessary distinction is Bigtable row-key distribution.
D:It does not correct the Bigtable row-key, cluster or data-access requirement. The case instead depends on Bigtable row-key distribution.
Question 272
A business intelligence team needs SQL analysis over terabytes of event records and doesn't require millisecond row-level transaction updates. Which service fits?
- Use Cloud SQL as a single-instance replacement for the requested large-scale analytics query.
- Store the dataset only as Cloud Storage objects and assume SQL aggregation happens automatically.
- Load the analytic dataset into BigQuery with an appropriate table design and query controls.
- Use Bigtable key/value reads for this SQL analytical reporting workflow.
Explanation
Correct Answer
C:BigQuery is designed for analytical SQL workloads rather than an operational transactional database role.
Incorrect Answers
A:The alternative does not address the query or storage behavior driving the workload. The necessary distinction is the requested BigQuery operation.
B:The alternative does not address the query or storage behavior driving the workload. The case instead depends on the requested BigQuery operation.
D:The alternative does not address the query or storage behavior driving the workload. It leaves the requested BigQuery operation unaddressed.
Question 273
A mobile app needs managed document storage with flexible structured records and automatic client data synchronization support. Which data model is most relevant?
- Create the required composite index for the query shape, then test after the index reaches its ready state.
- Change the instance display name as a replacement for modifying the actual machine or disk.
- Use Firestore with the appropriate document collections, indexes and security rules.
- Create a snapshot only and assume it immediately changes the live VM capacity.
Explanation
Correct Answer
C:Firestore is a managed document database rather than a VM-mounted filesystem or column-family store.
Incorrect Answers
A:The suggested action changes another component without meeting the main requirement. The case instead depends on the requested platform behavior.
B:Display names and labels identify resources; they do not enforce their runtime behavior. It leaves the requested platform behavior unaddressed.
D:Snapshots are recovery artifacts and do not directly reconfigure a running resource. The specific requirement concerns the requested platform behavior.
Question 274
A transactional application needs PostgreSQL compatibility but requires evaluation of AlloyDB-specific enterprise performance, read pools, and managed availability features before migration. Which service should architects assess?
- Cloud Storage Archive class, because changing object storage class provides PostgreSQL indexes and ACID transactions.
- Cloud Logging, because retained audit entries behave like a managed PostgreSQL data engine.
- AlloyDB for PostgreSQL, after validating its documented PostgreSQL compatibility, sizing, and regional features.
- Cloud CDN, because caching external requests replaces relational PostgreSQL storage without application changes.
Explanation
Correct Answer
C:AlloyDB for PostgreSQL, after validating its documented PostgreSQL compatibility, sizing, and regional features. AlloyDB is a managed PostgreSQL-compatible relational service with product-specific performance and availability features to evaluate for this workload.
Incorrect Answers
A:Cloud Storage Archive class, because changing object storage class provides PostgreSQL indexes and ACID transactions. Storage classes control object access and costs, not relational execution or PostgreSQL extension compatibility.
B:Cloud Logging, because retained audit entries behave like a managed PostgreSQL data engine. Cloud Logging stores telemetry entries and is not a transactional SQL database with application table and query support.
D:Cloud CDN, because caching external requests replaces relational PostgreSQL storage without application changes. Cloud CDN can cache eligible HTTP responses but cannot replace the transactional database that maintains the application's relational state.
Question 275
Independent producers and consumers need asynchronous delivery of application events with independent scaling. Which product is appropriate?
- Send events directly to a single synchronous HTTP endpoint and couple every producer to consumer availability.
- Store messages in a Cloud Storage bucket and expect it to implement Pub/Sub acknowledgements.
- Publish messages to Pub/Sub topics and connect authorized subscriptions to consumers.
- Use Cloud Logging as the application event broker and assume it provides the needed subscription delivery.
Explanation
Correct Answer
C:Pub/Sub decouples event producers and subscribers without requiring them to share a synchronous database transaction.
Incorrect Answers
A:The suggested action changes another component without meeting the main requirement. The specific requirement concerns message delivery or subscription behavior.
B:The suggested action changes another component without meeting the main requirement. The necessary distinction is message delivery or subscription behavior.
D:Cloud Logging records telemetry, rather than directly enforcing the requested runtime behavior. The case instead depends on message delivery or subscription behavior.
Question 276
An ETL system needs managed stream and batch transformations using Apache Beam semantics. Which data-processing service is designed for this?
- Run a one-time BigQuery query and assume it provides a continuously managed streaming pipeline.
- Run the eligible pipelines in Dataflow with the required inputs, outputs and monitoring.
- Increase the source bucket storage class rather than correcting the Beam transform or pipeline.
- Grant organization-wide Editor to pipeline workers instead of identifying the required access.
Explanation
Correct Answer
B:Dataflow manages execution of Apache Beam-based batch and streaming pipelines.
Incorrect Answers
A:The suggested action changes another component without meeting the main requirement. The necessary distinction is the managed data-transformation pipeline.
C:Storage class changes cost and access behavior, not the required policy by itself. The case instead depends on the managed data-transformation pipeline.
D:The suggested action changes another component without meeting the main requirement. It leaves the managed data-transformation pipeline unaddressed.
Question 277
An organization has existing Apache Kafka clients and protocols and wants Google-managed broker operations. What service should it evaluate?
- Use Google Cloud Managed Service for Apache Kafka after validating cluster sizing and client connectivity.
- Give anonymous users Storage Admin as a shortcut to object access management.
- Use a Filestore NFS share as though it implemented Cloud Storage object versioning and generation semantics.
- Use a VM-local Persistent Disk as the only copy of the object data without the required bucket controls.
Explanation
Correct Answer
A:A managed Kafka service preserves supported Kafka interfaces rather than forcing an unrelated messaging protocol.
Incorrect Answers
B:Broad Storage Admin access introduces privileges beyond the stated task. The case instead depends on the object's required data or access behavior.
C:Object versioning and the requested retention or deletion safeguards are separate mechanisms. It leaves the object's required data or access behavior unaddressed.
D:The option does not enforce the required object location, access or lifecycle behavior. The specific requirement concerns the object's required data or access behavior.
Question 278
A service needs a low-latency in-memory cache for frequently recomputed session lookups rather than a durable analytics warehouse. Which service fits?
- Choose an appropriate Memorystore in-memory data engine and sizing tier.
- Use a VM-local Persistent Disk as the only copy of the object data without the required bucket controls.
- Give anonymous users Storage Admin as a shortcut to object access management.
- Use a Filestore NFS share as though it implemented Cloud Storage object versioning and generation semantics.
Explanation
Correct Answer
A:Memorystore provides managed in-memory caching and compatible data-store engines for low-latency lookups.
Incorrect Answers
B:The option does not enforce the required object location, access or lifecycle behavior. It leaves the requested latency measurement unaddressed.
C:Broad Storage Admin access introduces privileges beyond the stated task. The specific requirement concerns the requested latency measurement.
D:Object versioning and the requested retention or deletion safeguards are separate mechanisms. The necessary distinction is the requested latency measurement.
Question 279
A team creates Bigtable row keys with an increasing timestamp prefix and notices persistent hot tablets under sequential writes. What design should it revisit?
- Distribute write hotspots with an appropriate row-key salting or ordering strategy tied to query patterns.
- Increase Bigtable node count while keeping the monotonically increasing row-key prefix unchanged.
- Rebuild only the client VM image while leaving the Bigtable table configuration unchanged.
- Change the Bigtable cluster display name as a substitute for the required data operation.
Explanation
Correct Answer
A:Sequential row-key ranges can concentrate Bigtable write traffic on a narrow set of tablets.
Incorrect Answers
B:More nodes may add capacity, but sequential keys can keep writes concentrated on hot tablets.
C:It does not correct the Bigtable row-key, cluster or data-access requirement. The necessary distinction is Bigtable row-key distribution. The appropriate action here is: Distribute write hotspots with an appropriate row-key salting or ordering strategy tied to query patterns.
D:Display names and labels identify resources; they do not enforce their runtime behavior. The case instead depends on Bigtable row-key distribution.
Question 280
A globally distributed transactional schema uses a monotonically increasing leading primary-key segment and suffers write hotspots. What should be redesigned?
- Select an eligible Spanner multi-region instance configuration with an approved replication topology and data residency.
- Use a backup as the only solution to a live Spanner query or scaling constraint.
- Change only the Spanner instance label, leaving placement and schema untouched.
- Use a Spanner key scheme that distributes sequential writes while preserving required query access patterns.
Explanation
Correct Answer
D:Sequential leading keys can produce write hotspots in distributed databases even when the service automatically shards storage.
Incorrect Answers
A:Placement changes do not automatically implement the required resource-level feature. The necessary distinction is distributed write-key placement.
B:It does not supply the specified transactional schema or consistency behavior. The case instead depends on distributed write-key placement.
C:It does not supply the specified transactional schema or consistency behavior. It leaves distributed write-key placement unaddressed.