PDE Practice Exam 52 — 20 Questions

Question 1

A company uses Cloud Storage to store IoT sensor data in JSON format. The data is ingested using a Cloud Function triggered by Cloud Storage events. They notice that when many files are uploaded simultaneously, some files are not processed and the Cloud Function logs show 'function execution timeout'. What is the most likely cause and solution?

Choose an answer to begin — your selection is scored in the full session.

Accepted Answer

The Cloud Function's timeout is too short; increase timeout beyond 540 seconds.. The Cloud Function logs explicitly show 'function execution timeout', which indicates the function is exceeding its configured maximum runtime. The default Cloud Functions timeout is 60 seconds, and the maximum is 540 seconds (9 minutes). When many files are uploaded simultaneously, each function invocation may take longer due to increased processing load, causing timeouts. Increasing the timeout to the maximum of 540 seconds gives the function more time to complete processing, directly addressing the logged error.

Answer

The Cloud Function is not idempotent; implement idempotency.

Answer

The Cloud Storage event notification is unreliable; switch to Pub/Sub notifications.

Answer

The Cloud Function has too few instances; increase max instances.

Question 2

Refer to the exhibit. A team uses this Cloud Build configuration to deploy a service to Cloud Run. The deployment step fails with a 'Permission denied' error. What is the most likely cause?

Accepted Answer

The Cloud Build service account does not have the Cloud Run Admin role.. The Cloud Build service account (typically the default compute engine service account or a user-specified service account) must have the Cloud Run Admin role (roles/run.admin) to deploy services to Cloud Run. Without this IAM permission, the deployment step fails with a 'Permission denied' error, even if the build itself succeeds. The error occurs because Cloud Build attempts to call the Cloud Run Admin API (run.googleapis.com) to create or update the service, and the service account lacks the required authorization.

Answer

The Dockerfile is missing from the repository.

Answer

The Docker image tag is missing or malformed.

Answer

The region 'us-central1' is incorrect for Cloud Run.

Question 3

You are building a streaming pipeline to ingest real-time clickstream data from a website into BigQuery for immediate analysis. The data must be available in BigQuery within seconds and you need to handle late-arriving data (e.g., browser offline events) that may arrive hours later. Which approach should you use?

Accepted Answer

Use Pub/Sub with Dataflow, writing to BigQuery using the Storage Write API in committed mode.. Pub/Sub provides a scalable, durable ingestion layer for real-time clickstream data, and Dataflow can handle late-arriving data via its built-in watermark and trigger mechanisms. The Storage Write API in committed mode ensures exactly-once semantics and low-latency writes to BigQuery, meeting the sub-second availability requirement while preserving data consistency for delayed events.

Answer

Use Cloud Logging to capture logs and export to BigQuery via a sink.

Answer

Use Pub/Sub with Cloud Functions, writing each event directly via BigQuery legacy streaming inserts.

Answer

Use Datastream to stream clickstream data from Cloud SQL to BigQuery.

Question 4

Which TWO actions should be taken to optimize a Dataflow streaming pipeline that is experiencing high system lag and backpressure? (Choose two.)

Accepted Answer

Increase the number of worker threads by adjusting the streaming worker's parallelism hint.. Increasing the parallelism hint allows each worker to process more bundles concurrently, which can reduce backpressure by improving throughput without adding more workers. Option C is correct because enabling autoscaling and increasing the maximum number of workers allows the pipeline to dynamically scale out to handle increased load, directly mitigating high system lag and backpressure.

Answer

Use a higher memory machine type for all workers.

Answer

Reduce the number of workers to decrease cost.

Answer

Set maxNumWorkers to 1 to force single-worker processing.

Question 5

A company is migrating their on-premises Apache Spark jobs to Dataproc. They want to minimize code changes and take advantage of serverless infrastructure. Which Dataproc feature should they use?

Accepted Answer

Dataproc Serverless Spark. Dataproc Serverless Spark is the correct choice because it allows the company to run Spark workloads without provisioning or managing clusters, minimizing code changes by using the same Spark APIs and libraries. This serverless infrastructure automatically scales resources and handles failures, aligning with the goal of reducing operational overhead while maintaining compatibility with existing Spark jobs.

Answer

Dataproc clusters with preemptible VMs

Answer

Dataproc Workflow Templates

Answer

Dataproc Jobs API with custom machine types

Question 6

A company is designing a real-time analytics pipeline using Pub/Sub and Dataflow. They need to ensure exactly-once processing and handle late-arriving data. Which two configurations should they implement? (Choose TWO.)

Accepted Answer

Enable exactly-once delivery on Pub/Sub subscription. To achieve exactly-once semantics, enable exactly-once delivery on the Pub/Sub subscription (option B). To handle late-arriving data, use a fixed window with allowed lateness and a trigger (option D). This configuration ensures that late data is still processed within the allowed lateness period, while the trigger allows early results to be emitted before the window closes.

Answer

Set up a global window with no triggers

Answer

Use Dataflow's default at-least-once mode

Answer

Write all data to Cloud Storage and then batch load to BigQuery

Question 7

A team is designing an event-driven data pipeline. They need to process messages from Cloud Pub/Sub, transform them, and write to BigQuery. The messages have variable volume and spikes. What is the best serverless compute option for this workload?

Accepted Answer

Cloud Dataflow with a streaming pipeline. Cloud Dataflow with a streaming pipeline is the best serverless compute option because it is purpose-built for unbounded, variable-volume data streams from Pub/Sub and provides exactly-once processing semantics, auto-scaling, and built-in BigQuery sink integration via the Beam SDK. Unlike simpler compute options, Dataflow handles backpressure, windowing, and state management natively, making it ideal for spikes and high-throughput transformations without manual scaling or idempotency concerns.

Answer

Cloud Functions triggered by Pub/Sub

Answer

Compute Engine with a Pub/Sub client library

Answer

Cloud Run invoked via Eventarc

Question 8

A company has a data lake on Cloud Storage with raw data in the 'raw' bucket, curated data in 'curated', and processed data in 'processed'. They want to implement lifecycle management to reduce costs. Which TWO actions should they take? (Choose 2)

Accepted Answer

Set a lifecycle rule to change storage class from Standard to Nearline after 30 days for the 'raw' bucket.. Setting a lifecycle rule to change storage class from Standard to Nearline after 30 days for the 'raw' bucket reduces costs while maintaining quick access to frequently needed raw data. For 'curated' and 'processed' buckets, deleting objects older than 365 days is appropriate because these datasets are typically intermediate or final and can be removed after a retention period. Option B (object versioning) does not automatically delete older versions; it preserves them. Option C (partition expiration on BigQuery) applies to BigQuery tables, not Cloud Storage objects. Option E (changing to Archive after 30 days) is less cost-effective than Nearline for raw data that may still be accessed occasionally.

Answer

Enable object versioning on all buckets to automatically delete older versions.

Answer

Set a partition expiration on BigQuery tables that reference data in the 'processed' bucket.

Answer

Set a lifecycle rule to change storage class from Standard to Archive after 30 days for the 'raw' bucket.

Question 9

A company runs BigQuery workloads with varying demand. They want to use flat-rate pricing with baseline slots and the ability to burst during peak times. Which TWO actions should they take? (Choose 2)

Accepted Answer

Use flex slots for short-term bursts. To achieve flat-rate pricing with baseline slots and bursting, purchase committed use reservations for the baseline (option D) to get a discount on consistent usage, and use flex slots for short-term bursts (option B) which are also flat-rate. Option E is incorrect because autoscaling with reservations can use idle committed slots for bursts, but it may still fall back to on-demand pricing if no idle slots are available, so it does not ensure that all bursts are covered by flat-rate pricing. Flex slots (option B) guarantee flat-rate bursting.

Answer

Use on-demand pricing for bursting

Answer

Set a maximum number of slots per query

Answer

Create a reservation with baseline + autoscaling slots

Question 10

An e-commerce company uses Vertex AI to serve a real-time personalization model. The model is updated daily via a retraining pipeline that uploads a new version to the same endpoint. Recently, after a model update, the online prediction responses have been returning anomalous results (e.g., recommending irrelevant products). The previous version performed well. The team suspects that the new model is undercooked or has a bug. They have already checked the training code and the pipeline logs, which show no errors. The pipeline deploys the new model version to the endpoint by updating the traffic split to route 100% of traffic to the new version. Which course of action should the team take to quickly mitigate the issue while diagnosing the root cause? A) Roll back the endpoint to the previous model version by setting traffic split to 0% for the new version. B) Delete the current endpoint and recreate it with the previous model version. C) Tweak the training hyperparameters and retrain immediately. D) Increase the number of replicas on the endpoint to handle load.

Accepted Answer

Roll back the endpoint to the previous model version by setting traffic split to 0% for the new version.. Rolling back traffic to the previous known-good version immediately restores correct predictions, while the team investigates the new model. Option A (tweaking hyperparameters) takes time and may not fix the bug. Option B (increasing replicas) does not address the incorrect model output. Option C (deleting endpoint) is excessive and causes downtime.

Answer

Tweak the training hyperparameters and retrain immediately.

Answer

Increase the number of replicas on the endpoint to handle load.

Answer

Delete the current endpoint and recreate it with the previous model version.

Question 11

A team uses Vertex AI Feature Store for real-time features. They notice that features are frequently missing during prediction serving. What is the best practice to handle missing features?

Accepted Answer

Use a default value in the feature store definition. Vertex AI Feature Store allows you to define a default value for each feature at the time of feature store creation or feature definition. When a feature value is missing during serving, the feature store automatically returns this default value instead of failing or returning null. This ensures that the serving function always receives a valid feature value without requiring custom imputation logic or model retraining.

Answer

Retrain the model to handle missing values

Answer

Impute missing values in the serving function

Answer

Drop the prediction request

Question 12

A retail company uses BigQuery to store sales data and wants to forecast weekly demand for the next 8 weeks using historical data from the past 2 years. They need to account for seasonality and holidays. Which BigQuery ML model type and configuration is most appropriate?

Accepted Answer

ARIMA_PLUS with holiday_region parameter. BigQuery ML's ARIMA_PLUS model is designed for time-series forecasting, automatically detecting seasonality and handling holiday effects via the holiday_region parameter. Linear regression would require manual feature engineering for time components. Time-series DECOMPOSE is not a model type. Boosted trees are not natively time-series aware without feature engineering.

Answer

Boosted tree classifier

Answer

Linear regression with engineered time features

Answer

Time-series DECOMPOSE model

Question 13

Drag and drop the steps to set up a BigQuery dataset with a scheduled query into the correct order.

Accepted Answer

Create a BigQuery dataset, then write the SQL query, then configure the scheduled query.. Setting up a BigQuery dataset with a scheduled query involves three logical steps in order: first create a dataset to store your data, then write the SQL query that transforms or loads the data, and finally configure the scheduled query to run at specified intervals. The dataset must exist before you can reference it in the query or schedule, and the query must be defined before you can schedule it. Reversing these steps leads to errors because each step depends on the previous one.

Answer

Write the SQL query, then create the BigQuery dataset, then configure the scheduled query.

Answer

Configure the scheduled query, then create the BigQuery dataset, then write the SQL query.

Answer

Create the BigQuery dataset, then configure the scheduled query, then write the SQL query.

Question 14

You need to create a Cloud Storage bucket for a data lake that will store raw ingested data. The data must be immutable and cannot be deleted or overwritten for a compliance period of 5 years. Which feature should you enable?

Accepted Answer

Bucket Lock with a retention policy of 5 years. Bucket Lock with a retention policy enforces a minimum retention period on all objects in the bucket. During the retention period, objects cannot be deleted or overwritten. This is exactly for compliance needs.

Answer

Object Versioning

Answer

Lifecycle rules to delete objects after 5 years

Answer

Object Lock with governance mode

Question 15

You are monitoring a Dataproc cluster and notice that the cluster utilisation is high, but jobs are running slowly. The cluster uses preemptible workers for cost savings. What is the most likely cause of the performance degradation?

Accepted Answer

The preemptible workers are being preempted frequently, causing task retries and slowdowns.. Preemptible workers can be terminated at any time, causing job restarts and slower execution. The high utilisation indicates many workers are preempted, leading to recomputation.

Answer

The primary workers are using standard disks instead of SSDs.

Answer

The cluster is under-provisioned; increase the number of preemptible workers.

Answer

The cluster is using an older image version; upgrade to the latest.

Question 16

A company runs Apache Kafka on Dataproc for real-time event streaming. They want to archive the Kafka topics to Cloud Storage for long-term retention and later analysis in BigQuery. Which approach is the most cost-effective and operationally simple?

Accepted Answer

Use Kafka Connect with the GCS Sink Connector to write directly to Cloud Storage. Kafka Connect with the GCS Sink Connector is purpose-built for exactly this use case: it directly streams Kafka topics to Cloud Storage in Avro, Parquet, or JSON format without requiring intermediate processing clusters or services. This approach minimizes operational overhead (no Spark or Dataflow jobs to manage) and is cost-effective since it runs as a lightweight connector within the existing Kafka ecosystem, leveraging Dataproc's managed Kafka cluster.

Answer

Use Apache Spark streaming on Dataproc to read from Kafka and write to GCS

Answer

Use Kafka MirrorMaker to replicate topics to a second cluster that writes to GCS

Answer

Use the Pub/Sub connector to publish Kafka messages to Pub/Sub, then a Dataflow job to write to GCS

Question 17

A gaming company uses Avro schemas for its streaming event data. They anticipate adding new optional fields to events over time. They need to ensure backward compatibility so that existing pipelines continue to work. Which strategy should they adopt?

Accepted Answer

Use Avro with a schema registry that enforces backward-compatible changes. Avro, combined with a schema registry, allows schema evolution with backward compatibility. The registry enforces rules such as adding optional fields with defaults, ensuring that consumers using older schemas can still deserialize new data without breaking. This directly addresses the requirement for existing pipelines to continue working as new optional fields are added.

Answer

Use JSON instead of Avro and ignore unknown fields

Answer

Use Protocol Buffers with breaking changes

Answer

Use FlatBuffers for performance

Question 18

A team monitors a deployed Vertex AI model and notices an increasing number of prediction errors with status code 413 (Request Entity Too Large). Which TWO actions should they consider to resolve this issue?

Accepted Answer

Implement client-side pre-processing to compress or downsample input data. Status code 413 indicates the HTTP request payload exceeds the server's size limit. Implementing client-side pre-processing to compress or downsample input data reduces the payload size before it reaches the Vertex AI endpoint, directly addressing the root cause. This approach is efficient because it shifts the computational burden to the client and avoids hitting the server-imposed request size cap, which is typically 1.5 MB for online predictions in Vertex AI.

Answer

Switch the model to batch prediction to handle large payloads offline

Answer

Increase the number of replicas to handle load

Answer

Decrease the machine type to reduce resource consumption

Question 19

A company uses BigQuery partitioned tables with daily partitions for log data. They want to automatically delete partitions older than 90 days and ensure that current month data is in a specific dataset with a set expiration. Which TWO actions should they take? (Choose 2)

Accepted Answer

Set partition expiration to 90 days on the table. To automatically delete partitions older than 90 days, the table must be partitioned (option C creates a partitioned table by ingestion time) and have partition expiration set to 90 days (option A). Option D also sets partition expiration, but it's functionally identical to A. Since the question asks for two distinct actions, the correct choices are A and C, which together achieve the goal.

Answer

Cluster the table on a timestamp column

Answer

Set the table's default partition expiration to 90 days

Answer

Set a lifecycle management rule on Cloud Storage to delete objects older than 90 days

Question 20

A company uses BigQuery Omni to query data stored in AWS S3. They need to join this data with data in BigQuery (GCP). The dataset in AWS is large (10 TB) and frequently updated. Which approach minimizes data movement and cost?

Accepted Answer

Use BigQuery Omni cross-cloud join with the query processed in AWS region. BigQuery Omni supports cross-cloud joins by running the query on the cloud where the data resides, using the cross-cloud join feature to pull only the necessary rows from the other cloud.

Answer

Create a materialized view in BigQuery that includes the AWS data

Answer

Copy the AWS data into BigQuery storage using a scheduled transfer, then join

Answer

Use a federated query from BigQuery to read the AWS data directly and join in BigQuery