Google Associate Cloud Engineer Practice Test 46

Question 450

An administrator wants to limit data loss after an accidental change by restoring to a specific supported timestamp. Which approach should be selected?

  1. Use Cloud SQL point-in-time recovery when configured and supported
  2. Restore the Cloud SQL backup to an appropriate target instance
  3. Use Cloud SQL high availability with a configured standby
  4. Create a supported Cloud SQL read replica

Explanation

Correct Answer

A:

PITR uses available transaction logs and backup prerequisites to restore to a chosen time. It is not available merely because an ordinary VM disk snapshot exists.

Incorrect Answers

B:

Its purpose is to restore a managed database backup. That distinction matters: the scenario calls for an action to undo an accidental database operation, not this separate configuration task.

C:

Administrators use it to use database instance HA failover, rather than to undo an accidental database operation. Applying it would change the wrong component of this deployment.

D:

The result of this choice would be to offload relational read operations. The requested outcome, to undo an accidental database operation, is handled at a different layer or through another operation.

Question 451

A Cloud SQL read replica has started serving stale query results because replication is lagging. The primary instance is healthy. What should the operator inspect first?

  1. Assume a high-availability standby is the active read replica and query it instead
  2. Change the Cloud SQL maintenance window and assume all replication lag disappears
  3. Check replica-lag metrics and replication status, then investigate write load and replica capacity
  4. Use point-in-time recovery to restore the primary to an earlier timestamp

Explanation

Correct Answer

C:

Read replicas apply changes asynchronously, so replication lag can lead to stale reads. Replication metrics and workload capacity help identify whether the replica can keep up with the primary.

Incorrect Answers

A:

An HA standby is for availability and failover, not a normal independently addressable read-replica endpoint. It does not remove asynchronous lag on a separate read replica.

B:

A maintenance window schedules supported maintenance timing. It does not by itself correct an undersized replica or a write workload that outpaces replication.

D:

Point-in-time recovery restores a database to a selected earlier state. It is a recovery tool, not an appropriate first response to lag on a healthy primary's read replica.

Question 452

A database needs a separate read endpoint to offload suitable SELECT traffic. What is the best next step?

  1. Use Cloud SQL high availability with a configured standby
  2. Create a supported Cloud SQL read replica
  3. Restore the Cloud SQL backup to an appropriate target instance
  4. Create a Spanner backup using the service's supported backup feature
  5. Use Cloud SQL point-in-time recovery when configured and supported

Explanation

Correct Answer

B:

Read replicas help scale reads and can have asynchronous replication lag. They are not a replacement for the primary write endpoint.

Incorrect Answers

A:

This setting would use database instance HA failover. That behavior does not establish the requested ability to reduce SELECT load on the primary; it addresses an adjacent concern instead.

C:

Its purpose is to restore a managed database backup. That distinction matters: the scenario calls for an action to reduce SELECT load on the primary, not this separate configuration task.

D:

This alternative addresses spanner and alloydb management by allowing a team to back up a Spanner database; The incident, however, requires the different action to reduce SELECT load on the primary, so its central configuration remains unchanged.

E:

This approach can recover relational data to a prior time. The requirement is to reduce SELECT load on the primary, so it controls a different part of the workflow and leaves that result unaddressed.

Question 453

An application administrator needs to run a transactional SQL query against a Cloud SQL instance. What should the administrator do?

  1. Submit a BigQuery SQL query job in the correct location
  2. Look up its BigQuery job status and errors
  3. Connect with an authorized SQL client or Cloud SQL connector
  4. Filter on the partition column for a partitioned table

Explanation

Correct Answer

C:

Cloud SQL exposes the selected relational engine to compatible clients. Network access, database credentials and IAM for supporting services must all be handled.

Incorrect Answers

A:

The result of this choice would be to execute warehouse SQL analytics. The requested outcome, to query an operational relational database, is handled at a different layer or through another operation.

B:

This mechanism is intended to inspect BigQuery query job state. It does not implement the required action to query an operational relational database, even though both are relevant to the same general service.

D:

The tool can prune analytical partitions in SQL, but it cannot be substituted for the capability needed to query an operational relational database. Its operational effect is not equivalent.

Question 454

An analyst needs to run an aggregate SELECT over a BigQuery table. Which choice provides the required behavior?

  1. Look up its BigQuery job status and errors
  2. Submit a BigQuery SQL query job in the correct location
  3. Connect with an authorized SQL client or Cloud SQL connector
  4. Filter on the partition column for a partitioned table
  5. Create a supported Cloud SQL read replica

Explanation

Correct Answer

B:

BigQuery executes SQL analytics through managed query jobs. Query costs and permissions differ from reading a single object from Cloud Storage.

Incorrect Answers

A:

This setting would inspect BigQuery query job state. That behavior does not establish the requested ability to aggregate BigQuery table data; it addresses an adjacent concern instead.

C:

Using this action would execute SQL against Cloud SQL, not aggregate BigQuery table data. Its function is different from the specific change described in the scenario.

D:

This choice is useful to prune analytical partitions in SQL. The stated task is to aggregate BigQuery table data, which requires a different mechanism or permission scope.

E:

The choice is aimed at a different purpose: to reduce SELECT load on the primary; Even if it can offload relational read operations, that would not accomplish the stated requirement to aggregate BigQuery table data.

Question 455

An operator must inspect a previously submitted BigQuery job that has not finished. Which approach should be selected?

  1. Connect with an authorized SQL client or Cloud SQL connector
  2. Look up its BigQuery job status and errors
  3. Submit a BigQuery SQL query job in the correct location
  4. Filter on the partition column for a partitioned table
  5. Restore the Cloud SQL backup to an appropriate target instance

Explanation

Correct Answer

B:

BigQuery job metadata reports state, statistics and failures. It separates query execution status from table existence.

Incorrect Answers

A:

This mechanism is intended to execute SQL against Cloud SQL. It does not implement the required action to diagnose a running or failed warehouse job, even though both are relevant to the same general service.

C:

Administrators use it to execute warehouse SQL analytics, rather than to diagnose a running or failed warehouse job. Applying it would change the wrong component of this deployment.

D:

The result of this choice would be to prune analytical partitions in SQL. The requested outcome, to diagnose a running or failed warehouse job, is handled at a different layer or through another operation.

E:

This alternative addresses cloud sql database maintenance by allowing a team to restore a managed database backup; The incident, however, requires the different action to diagnose a running or failed warehouse job, so its central configuration remains unchanged.

Question 456

A team needs to reduce unnecessary BigQuery bytes scanned by a query with a date predicate. Which method addresses the task?

  1. Grant a suitable Storage Object Viewer role on the specific bucket
  2. Connect with an authorized SQL client or Cloud SQL connector
  3. Submit a BigQuery SQL query job in the correct location
  4. Look up its BigQuery job status and errors
  5. Filter on the partition column for a partitioned table

Explanation

Correct Answer

E:

Partition pruning avoids scanning unrelated partitions where query planning allows. A row LIMIT alone often does not reduce scanned input data.

Incorrect Answers

A:

The tool can grant read access to objects in one bucket, but it cannot be substituted for the capability needed to control bytes scanned by queries. Its operational effect is not equivalent.

B:

Using this action would execute SQL against Cloud SQL, not control bytes scanned by queries. Its function is different from the specific change described in the scenario.

C:

It can execute warehouse SQL analytics; however, it does not provide the specific operation to control bytes scanned by queries. The two controls serve different purposes.

D:

This setting would inspect BigQuery query job state. That behavior does not establish the requested ability to control bytes scanned by queries; it addresses an adjacent concern instead.

Question 457

A team created the composite index recommended by Firestore for a large collection query, but the query still fails immediately after index creation while the console reports that the index is building. What should the operator do?

  1. Disable Firestore Security Rules so the index can finish building
  2. Increase the memory of a Compute Engine VM that is not running the managed Firestore service
  3. Wait until the Firestore index reports Ready, then retry the indexed query
  4. Delete the index immediately and use its definition only as documentation

Explanation

Correct Answer

C:

A composite index may take time to build over existing documents. Queries requiring it cannot rely on the index until it reaches the ready state.

Incorrect Answers

A:

Security rules govern client authorization and do not control the background progress of a composite index build.

B:

Firestore index builds are managed by the service. Enlarging an unrelated VM does not complete the managed index operation.

D:

Deleting the still-building index removes the resource needed by the query. The query cannot use a nonexistent index.

Question 458

A web client must be prevented from reading other users' Firestore documents. What should the administrator do?

  1. Use an appropriate Firestore transaction or batched write based on read needs
  2. Use a supported Firestore managed export or backup operation
  3. Create the required Firestore composite index
  4. Update Firestore Security Rules to enforce the authenticated user's ownership

Explanation

Correct Answer

D:

Rules can use authentication context and document paths to authorize client SDK operations. Database IAM alone is not a substitute for client access rules.

Incorrect Answers

A:

The tool can commit related document changes atomically, but it cannot be substituted for the capability needed to restrict client document reads. Its operational effect is not equivalent.

B:

This mechanism is intended to back up Firestore document data. It does not implement the required action to restrict client document reads, even though both are relevant to the same general service.

C:

Select this when the goal is to resolve a Firestore missing-index error. Here the decisive requirement is to restrict client document reads, and that outcome is not achieved by this action.

Question 459

A Firestore application receives an index-required error for one filtered and ordered query. The same query also returns permission errors for some user contexts. Which THREE independent areas should developers review? (Choose THREE.)

  1. Security rules or IAM authorization applying to the actual calling context
  2. The monthly Cloud Billing budget email recipients as the mechanism enforcing document ACLs
  3. The exact query filters and orderBy clauses used by the application
  4. A new VPC peering connection as a substitute for every Firestore query index
  5. The composite index suggested for the query's supported filter and ordering combination

Explanation

Correct Answers

A:

An index does not grant document-read permission. Different users or server-side principals may still be denied by the rules or resource IAM path governing their calls.

C:

Query structure determines index requirements and may affect which documents can be returned. Developers must confirm the application is issuing the intended valid query.

E:

Some Firestore queries require an appropriate composite index. The error typically identifies the index requirement, which is separate from data access authorization.

Incorrect Answers

B:

Cost alert recipients do not participate in Firestore document authorization. Changing who receives a billing email cannot fix index errors or read permissions.

D:

VPC Peering concerns network connectivity. It does not define required Firestore composite indexes or bypass database authorization checks.

Question 460

An application expects a multi-document update to be all-or-nothing. Which approach should be selected?

  1. Create the required Firestore composite index
  2. Update Firestore Security Rules to enforce the authenticated user's ownership
  3. Configure AlloyDB backup and recovery operations for the cluster
  4. Use an appropriate Firestore transaction or batched write based on read needs
  5. Use a supported Firestore managed export or backup operation

Explanation

Correct Answer

D:

Firestore supports atomic operations across supported documents. Transactions support read-modify-write patterns with retry semantics.

Incorrect Answers

A:

Select this when the goal is to resolve a Firestore missing-index error. Here the decisive requirement is to preserve consistency across document updates, and that outcome is not achieved by this action.

B:

This mechanism is intended to apply per-user Firestore data access. It does not implement the required action to preserve consistency across document updates, even though both are relevant to the same general service.

C:

Choosing this would protect AlloyDB with managed backups, an appropriate action for spanner and alloydb management operations; It cannot take the place of the separate step needed to preserve consistency across document updates in this specific situation.

E:

The result of this choice would be to back up Firestore document data. The requested outcome, to preserve consistency across document updates, is handled at a different layer or through another operation.

Associate Cloud Engineer Practice Tests