GCP-ADP Practice Test 21 — 15 Questions

Question 1

You need to track who accessed which GCS bucket. Which service provides this audit trail?

Accepted Answer

Cloud Audit Logs. Cloud Audit Logs is the correct answer because it automatically records all administrative and data access activities on GCS buckets, including who accessed what, when, and from where. Specifically, Admin Activity logs capture bucket configuration changes, while Data Access logs (when enabled) record object reads and writes. This provides the audit trail required to track access to GCS buckets.

Answer

Cloud Security Command Center

Answer

Cloud Monitoring

Answer

Cloud Trace

Question 2

Your organization stores sensitive customer data in a Cloud Storage bucket. You need to ensure that data is automatically classified and tagged as 'PII' when it is written, and that the tags are available for audit and access control decisions. You want to minimize manual effort and use Google Cloud native tooling. What should you do?

Accepted Answer

Enable Sensitive Data Protection (DLP) discovery on the bucket and configure a discovery scan with a schedule; then publish the results to Data Catalog.. Sensitive Data Protection discovery is the native Google Cloud service for automatically profiling and classifying data in Cloud Storage. It can detect PII infoTypes, run on a schedule, and publish findings as tags in Data Catalog. This satisfies the need for automatic classification with minimal manual effort and integrates with audit and access control.

Answer

Create a Cloud Function that triggers on object finalize events, calls the DLP API to inspect the object, and writes custom tags to a separate Firestore database.

Answer

Use Dataplex Data Quality tasks to scan the bucket and automatically apply PII tags based on predefined rules.

Answer

Configure a VPC Service Controls perimeter around the bucket and enable Access Transparency logs to classify the data.

Question 3

What does 'Data Lineage' show you in Dataplex?

Accepted Answer

The path the data took through the system. Data Lineage in Dataplex shows the path data takes through the system — from source to destination, including transformations and dependencies. It provides a visual and programmatic view of how data flows across datasets, tables, and processes, enabling impact analysis and root-cause debugging.

Answer

The number of users accessing the data

Answer

The cost of storing data

Answer

The database schema version

Question 4

Your team ingests JSON event files into Cloud Storage, and you must load them into BigQuery daily. The files contain a field named `user_info` that holds a nested JSON object with varying keys. You want BigQuery to store this field as a native STRUCT so analysts can query subfields with dot notation. Which schema type should you define for the `user_info` column when creating the table?

Accepted Answer

RECORD. BigQuery represents nested structures with the RECORD type, which is equivalent to a STRUCT. When you load JSON where a field contains an object, declaring that column as RECORD lets BigQuery preserve the nested keys and types, and analysts can query subfields with dot notation. STRING, BYTES, and GEOGRAPHY do not provide native nested-field access or schema enforcement.

Answer

BYTES

Answer

GEOGRAPHY

Answer

STRING

Question 5

A data engineer needs a lightweight, serverless orchestrator on Google Cloud to run a sequence of HTTP calls to internal microservices, retry the calls a few times on failure, and expose the whole sequence as a single REST endpoint. The workflow must not require provisioning any cluster or VM. Which Google Cloud service should be used?

Accepted Answer

Workflows. Workflows is the serverless orchestration product on Google Cloud and is purpose-built for sequences of steps such as HTTP calls with retries and error handling. It requires no cluster or VM, and each workflow can be executed through a REST endpoint. Cloud Composer is heavier and cluster-based, Cloud Scheduler only triggers jobs, and Dataflow is a data processing runner rather than a service orchestrator.

Answer

Dataflow

Answer

Cloud Composer

Answer

Cloud Scheduler

Question 6

You are designing a Cloud Composer 2 DAG that must run a task only after a daily file has arrived in a Cloud Storage bucket. You want the DAG to avoid failing prematurely and to avoid consuming a worker slot while waiting. Which two approaches should you use? (Choose two.)

Accepted Answer

Set the sensor's mode to reschedule so the task releases its worker slot between pokes.. Waiting for an external file is a sensor use case. GCSObjectExistenceSensor checks for the object and succeeds only when it appears, and running the sensor with mode set to reschedule frees the worker between pokes so a long wait does not occupy a slot. Bash-based polling abuses retries, a fixed delay does not reflect file arrival, and catchup is unrelated to external event waiting.

Answer

Configure the DAG's catchup parameter to True so missed runs backfill until the file appears.

Answer

Wrap a gsutil ls command in a BashOperator and rely on its non-zero exit code to retry the task until the file appears.

Answer

Use a TimeDeltaSensor set to 24 hours so the task always runs exactly one day after the DAG starts.

Question 7

Which TWO of the following are key benefits of using Dataflow?

Accepted Answer

Fully managed. Option A is correct because Dataflow is a fully managed service: Google handles provisioning, scaling, and maintenance of the underlying Compute Engine worker VMs, so you don't manage clusters or infrastructure yourself. Option C is correct because Dataflow's Apache Beam-based programming model natively supports both batch and streaming pipelines, using the same code with different runners and windowing/triggering semantics. The unmarked options do not belong: hardware management (B) is handled by Google as part of the managed service, not a benefit you perform; automatic data entry (D) is not a Dataflow feature since Dataflow processes data rather than capturing or entering it; and low-latency SQL queries (E) describe an interactive query engine such as BigQuery, not Dataflow's pipeline execution model.

Answer

Hardware management

Answer

Automatic data entry

Answer

Low-latency SQL queries

Question 8

You are processing streaming data in Pub/Sub and need to archive every message into GCS without writing custom code. What should you use?

Accepted Answer

Pub/Sub Cloud Storage subscription. A Pub/Sub Cloud Storage subscription is a built-in subscription type that writes messages directly to a GCS bucket without any custom code. You configure the subscription with a bucket, filename prefix/suffix, and format, and Pub/Sub handles batching and delivery. This satisfies the requirement of archiving every message to GCS with no custom code.

Answer

Dataflow

Answer

BigQuery Data Transfer Service

Answer

Cloud Functions

Question 9

A data analyst is building a Looker Studio report connected to a BigQuery table. The report needs to display data for the last 30 days, but the underlying table is updated daily at midnight. Which approach ensures the report always shows the most recent 30 days without manual intervention?

Accepted Answer

Create a date range dimension in the report and set a default date range of 'Last 30 days'.. The requirement is for a report that dynamically shows the last 30 days without manual intervention. Looker Studio's date range dimension feature allows setting a default range like 'Last 30 days', which recalculates based on the current date each time the report is viewed. This leverages built-in functionality and ensures the report stays current as the BigQuery table updates daily.

Answer

Add a filter to the report that explicitly selects the start and end dates for the current month.

Answer

Use a BigQuery scheduled query to materialize the last 30 days into a new table and connect the report to that table.

Answer

Create a calculated field that subtracts 30 days from the current date and use it as a filter.

Question 10

You need to grant a data analyst read-only access to a specific BigQuery dataset. The analyst should be able to query tables but must not be able to modify data or view other datasets in the project. Which IAM role should you grant on the dataset?

Accepted Answer

roles/bigquery.dataViewer. Granting roles/bigquery.dataViewer on the dataset provides read-only access to the data and metadata, allowing the analyst to query tables without modification rights. It is the least privilege role that meets the requirement and can be scoped to a specific dataset, preventing access to other datasets in the project.

Answer

roles/bigquery.dataEditor

Answer

roles/bigquery.user

Answer

roles/bigquery.metadataViewer

Question 11

You are designing a BigQuery table for a fleet of IoT devices that emit one reading per second. The table will hold years of data, and most queries filter by device identifier and a time range. You want to reduce the bytes scanned by these queries and control long-term storage cost. (Choose two.)

Accepted Answer

Cluster the table on the device identifier column.. Partitioning by the event timestamp and clustering by device identifier align the physical layout of the table with the dominant query pattern of filtering by device and time range. Together they let BigQuery prune partitions and skip blocks, cutting bytes scanned and cost, while still retaining years of data.

Answer

Enable a customer-managed encryption key on the table to reduce storage cost.

Answer

Create a materialized view that aggregates readings per device per hour and query only the view.

Answer

Set the table's default expiration to 30 days so older data is deleted automatically.

Question 12

Which TWO are common causes of Dataflow pipeline failures?

Accepted Answer

Bad input data. Bad input data (B) is a common cause of Dataflow pipeline failures because malformed records, schema mismatches, or unexpected nulls can trigger exceptions in DoFn processing or cause parse errors in sources like BigQuery or Pub/Sub, failing the pipeline. Incorrect configuration (C) is also a frequent cause, as issues such as wrong worker machine types, insufficient maxNumWorkers, missing tempLocation/stagingLocation, or incorrect IAM roles prevent the job from launching or cause runtime errors. Too much storage (A) is not a typical failure cause since Dataflow scales with Cloud Storage and Persistent Disk, and storage capacity is generally elastic. Fast network speeds (D) are beneficial, not harmful, to pipeline execution. Cloud console downtime (E) does not stop Dataflow jobs, which run on managed Compute Engine workers independent of the console UI.

Answer

Too much storage

Answer

Fast network speeds

Answer

Cloud console downtime

Question 13

You need to migrate legacy data to a Dataplex Lake. Which THREE of the following should you consider for a successful migration? (Choose three)

Accepted Answer

Define a clear zone strategy. Option A is correct because a Dataplex Lake is organized into zones (raw, curated, and analytics) that map to data maturity and access patterns, so defining a clear zone strategy ensures legacy data lands in the appropriate zone with the right governance and lifecycle policies. Option B is correct because auditing existing IAM permissions on the source data is essential to map legacy access controls to Dataplex and Cloud IAM roles, preventing over-privileged or broken access after migration. Option E is correct because enabling Dataplex discovery automatically scans and inventories existing data assets (e.g., in Cloud Storage or BigQuery), giving you a catalog of what must be migrated and how it is structured. Option C is wrong because disabling data quality checks removes validation that migrated data is complete and accurate, which is a migration risk rather than a best practice. Option D is wrong because manually copying every file with scp is not a scalable or supported approach for migrating legacy data into a Dataplex Lake; managed transfer services or Dataplex/Cloud Storage tooling should be used instead.

Answer

Disable all data quality checks

Answer

Manually copy every file using scp

Question 14

You have an automated pipeline that deletes files in a GCS bucket every hour. Dataplex shows the asset as 'Out of sync'. Why?

Accepted Answer

Discovery needs to rescan to update metadata. Dataplex Discovery periodically crawls GCS buckets to extract metadata and update the Dataplex Catalog. When files are deleted outside of Dataplex (e.g., by an automated pipeline), the cataloged metadata becomes stale, causing the asset to be marked 'Out of sync'. The correct resolution is to trigger a new discovery scan, which re-crawls the bucket and synchronizes the metadata with the current state. Thus, option C is correct.

Answer

The lake is locked for maintenance

Answer

The user does not have permission to delete files

Answer

The file deletion is violating the Dataplex policy

Question 15

You need to audit who accessed sensitive data in a Dataplex Lake. Which service should you monitor to see access logs for these data assets?

Accepted Answer

Cloud Audit Logs. Cloud Audit Logs, specifically Data Access audit logs, record all access to Dataplex Lake data assets, including who accessed what and when. Dataplex integrates with Cloud Audit Logs to provide a unified audit trail for data access across its assets. This is the correct service to monitor for auditing sensitive data access.

Answer

Dataplex Task Logs

Answer

BigQuery Query Logs

Answer

Data Catalog history