Google CloudUpdated 2026Last reviewed

Google Professional Cloud Database Engineer (PCDE) Practice Test

Design, Migrate, and Run Databases on Google Cloud.

500 scenario-based practice questions covering all 4 PCDE exam sections: database design, multi-technology management, migration, and highly available deployment. Mapped to the official exam guide.

No credit card required · Upgrade to 500 questions for $11.99

500 Exam-Style Questions
7-Day Money-Back Guarantee

Is this for you?

You run production databases and want practice choosing between Cloud SQL, AlloyDB, Spanner, Bigtable, Firestore, and Memorystore when cost, consistency, and scale requirements conflict
You are planning a move to Google Cloud and want to test your migration plans: zero or near-zero downtime versus an extended outage, a fallback plan, and reverse replication back to the source
You want drills on setting RTO, RPO, and point-in-time recovery targets and matching them to scheduled backups, exports, and replicas
You troubleshoot performance and want to practice reading the signs of slow queries, locking, missing indexes, and resource contention from database vitals and audit logs

Not for you if:

If your work is building data pipelines and analytics rather than running operational databases, Professional Data Engineer matches your role better
If you want leaked real exam questions, we do not offer braindumps; every question here is written against the published PCDE exam guide

The Google Professional Cloud Database Engineer (PCDE) is a professional-level Google Cloud certification. The exam has 50–60 questions, runs 2 hours, costs $200, and requires a passing score of Not published by Google. CertSharp provides 500 exam-style PCDE practice questions — 30 free with no credit card — each with a detailed explanation.

Start 2–3 weeks before your exam for best results
Updated for 2026 exam blueprint
Detailed explanations for every question

Exam Overview

PCDE Exam Details

Know exactly what to expect on exam day

Questions

50–60

Duration

2 hours

Passing Score

Not published by Google

Exam Cost

$200

Compare this fee with other certifications in our exam cost comparison, or check your readiness with the free PCDE readiness quiz.

Coverage

Exam Domains Covered

Every question maps to an official PCDE domain

1

Design innovative, scalable, and highly available cloud database solutions (~32%)

2

Manage a solution that can span multiple database technologies (~25%)

3

Migrate data solutions (~23%)

4

Deploy scalable and highly available databases in Google Cloud (~20%)

From the exam guide

What the PCDE exam actually tests

PCDE is about operating databases under requirements: the guide repeatedly asks you to evaluate, justify, or recommend an option given workload, availability, and cost constraints. It covers the full life of a database, from capacity planning and service choice through monitoring, backup, migration, and highly available deployment.

Sizing and service selection

You size compute and storage from current workload metrics and future requirements, and weigh machine types and storage types against cost. You also separate SQL from NoSQL needs, including structured, semi-structured, unstructured, and vector data, and judge managed, self-managed, bare metal, and partner offerings.

Availability and recovery targets

The guide asks you to compare zonal, regional, and multi-regional deployments, set maintenance windows to fit application availability, and design for RTO, RPO, and PITR. Section 4 extends this to provisioning HA, deploying read replicas, setting up multi-regional replication, and testing your DR plan.

Connectivity, access, and day-2 operations

Expect questions on networking, encryption and key management, IAM for database access, managing database users, and when a session pooler is justified. Day-2 work includes monitoring quotas, alerting on errors and performance metrics, managing upgrades of Google-managed databases, and tracking SLAs and SLOs.

Migration planning and tooling

Section 3 (~23%) covers choosing the right migration tool for databases hosted outside Google Cloud, converting DDL and DML, and planning the cutover. You need to know when to accept downtime, how to keep a fallback path, and when to replicate from Google Cloud back to the source.

Why CertSharp

Built for Exam Day, Not Just Study Time

Study guides teach concepts. We prepare you for the actual PCDE exam.

1

Cloud SQL, AlloyDB, Spanner, Bigtable, Firestore, and Memorystore scenarios force the same service-selection tradeoffs the PCDE guide describes

2

Migration questions cover tool choice, near-zero downtime cutovers, fallback plans, and reverse replication

3

Every question carries an explanation that says why the right answer fits and why each distractor does not

4

Aligned to the current exam guide, including database support for generative AI and vector use cases

Free Sample Questions

PCDE Practice Questions (Free Sample)

Real PCDE exam-style questions from our 500-question Google Professional Cloud Database Engineer bank, each with the correct answer and a detailed explanation.

Question 1 — PCDE

A caching tier must keep running through a zonal outage without manual action. Which Memorystore for Redis configuration should be chosen?

  • A

    Basic tier with persistence enabled, which restarts the node in another healthy zone

  • B

    Basic tier with a read replica count of one to take over when the node fails

  • C

    Standard tier, which keeps a replica in another zone and fails over automatically

  • D

    Standard tier placed in a single zone with RDB snapshots exported to Cloud Storage

Correct answer: C — Explanation

Standard tier provides a replica in a different zone with automatic failover. Basic tier runs a single node without a replica or failover, and snapshots to Cloud Storage support recovery rather than automatic takeover.

  1. Question 2Design innovative, scalable, and highly available cloud database solutionsEasy

    In Memorystore for Redis, which tier provides automatic failover to a replica node?

    • A

      The Archive tier, which stores replicas for long term disaster recovery

    • B

      The Basic tier, which keeps a hot standby node in a second zone and fails over automatically in seconds

    • C

      The Development tier, which is optimized for continuous production availability

    • D

      The Standard tier, which maintains a replica in another zone and fails over automatically

    Show answer & explanation

    Correct answer: D — The Standard tier, which maintains a replica in another zone and fails over automatically

    Standard tier Memorystore for Redis uses a replica in a different zone with automatic failover. The Basic tier is a single node without replication, and the Development and Archive tiers do not exist as offered Memorystore tiers.

  2. Question 3Migrate data solutionsEasy

    A retailer must migrate a production MySQL database to Cloud SQL with minimal downtime. Which approach fits best?

    • A

      Cloud Composer DAG that runs a nightly full mysqldump and reloads the target instance each time

    • B

      Storage Transfer Service scheduled transfer of the MySQL data directory to a Cloud Storage bucket

    • C

      Database Migration Service continuous migration job using binary log replication from the source

    • D

      Transfer Appliance shipment of the source disks followed by a manual import on a new instance

    Show answer & explanation

    Correct answer: C — Database Migration Service continuous migration job using binary log replication from the source

    Database Migration Service performs an initial dump and then applies ongoing changes from the binary log, so cutover downtime is short. Storage Transfer Service moves objects, not live database files. Transfer Appliance is for bulk offline transfer and adds days of delay. A nightly dump reload leaves a large data gap and long downtime.

  3. Question 4Manage a solution that can span multiple database technologiesMedium

    A DBA must be notified when storage use on a Cloud SQL instance exceeds 85 percent so that capacity can be reviewed. What is the correct setup?

    • A

      Schedule a weekly manual review of the console storage chart and record the usage figure in a ticket

    • B

      Create a log sink that exports data access logs to BigQuery and query them each morning

    • C

      Create a Cloud Monitoring alerting policy on the disk utilization metric with a notification channel

    • D

      Enable Query Insights and wait for a storage warning

    Show answer & explanation

    Correct answer: C — Create a Cloud Monitoring alerting policy on the disk utilization metric with a notification channel

    An alerting policy on disk utilization notifies through email, chat, or paging channels when the threshold is crossed. Manual checks are slow, Query Insights targets query performance, and a log sink does not raise alerts on storage.

  4. Question 5Deploy scalable and highly available databases in Google CloudEasy

    After an AlloyDB zonal failover, what should application owners expect regarding connectivity?

    • A

      Open connections remain active and in-flight transactions continue without interruption

    • B

      The standby's IP replaces the primary's IP, so clients must be updated with the new address

    • C

      Clients must connect through a new regional endpoint that is created during the failover

    • D

      The primary instance IP stays the same, so clients reconnect without configuration changes

    Show answer & explanation

    Correct answer: D — The primary instance IP stays the same, so clients reconnect without configuration changes

    AlloyDB keeps the primary endpoint stable across failover, but connections are dropped and must be re-established. Client retry logic is still needed.

  5. Question 6Design innovative, scalable, and highly available cloud database solutionsMedium

    A retailer runs a MySQL workload that peaks at 12,000 transactions per second during sales. CPU averages 35 percent but spikes to 90 percent at peak. They want to size Cloud SQL without paying for idle capacity most of the year. What is the best approach?

    • A

      Size for the 90 percent peak value permanently and buy a three-year commitment to lock in the discounted rate

    • B

      Choose the smallest shared-core machine type and rely on burst credits during sales

    • C

      Use HDD storage to offset compute cost and keep the instance at its current size

    • D

      Size for the sustained load plus headroom, then scale vCPUs ahead of known sale events

    Show answer & explanation

    Correct answer: D — Size for the sustained load plus headroom, then scale vCPUs ahead of known sale events

    Right-sizing means covering the sustained baseline with headroom and temporarily scaling for predictable peaks. Permanently sizing for the peak with a long commitment pays for unused capacity most of the year. Shared-core machine types are not suitable for sustained production OLTP, and HDD does not lower compute demand.

  6. Question 7Migrate data solutionsMedium

    Part of a data set is already migrated to Cloud SQL, and analysts want to query it from BigQuery without building an ETL pipeline. What should be used?

    • A

      A BigQuery connection to Cloud SQL and the EXTERNAL_QUERY function in federated queries

    • B

      A Transfer Appliance that carries the Cloud SQL tables into BigQuery on removable disks

    • C

      A Datastream stream that writes the Cloud SQL tables into BigQuery as a destination database

    • D

      A Storage Transfer Service job that mounts the Cloud SQL instance as a source bucket

    Show answer & explanation

    Correct answer: A — A BigQuery connection to Cloud SQL and the EXTERNAL_QUERY function in federated queries

    Federated queries let BigQuery read Cloud SQL directly through a connection, which avoids ETL during a phased migration. Datastream would still create a copy, appliances are for offline transfer, and Cloud SQL is not a bucket source.

  7. Question 8Manage a solution that can span multiple database technologiesHard

    A Cloud SQL instance has HA enabled and PITR enabled. A bad deployment corrupts data at 14:05, and an operator notices at 14:40. What is the correct recovery plan?

    • A

      Restore the nightly backup from the previous day over the instance

    • B

      Promote a read replica, which keeps its own earlier version of the data from before the deployment

    • C

      Clone or restore to a timestamp just before 14:05, validate data, then redirect applications

    • D

      Fail over to the standby, since it contains the clean copy

    Show answer & explanation

    Correct answer: C — Clone or restore to a timestamp just before 14:05, validate data, then redirect applications

    PITR allows restoring to a time before the corruption, preserving most data. The standby and replicas replicate the corrupted changes. Restoring the nightly backup loses the day of valid changes.

  8. Question 9Deploy scalable and highly available databases in Google CloudMedium

    A DBA wants Spanner to scale between 1000 and 5000 processing units based on CPU and storage. Which command configures managed autoscaling?

    • A

      gcloud spanner instances update prod-spanner --autoscaling-min-processing-units=1000 --autoscaling-max-processing-units=5000 --autoscaling-high-priority-cpu-target=65

    • B

      gcloud spanner instances update prod-spanner --autoscale=true --min-units=1000 --max-units=5000 --target-cpu=65 --scaling-mode=managed

    • C

      gcloud spanner instances update prod-spanner --scaling-policy=cpu --target=65 --range=1000-5000 --scaling-mode=managed

    • D

      gcloud spanner instances update prod-spanner --min-nodes=1 --max-nodes=5 --cpu-target-percent=65 --scaling-mode=managed

    Show answer & explanation

    Correct answer: A — gcloud spanner instances update prod-spanner --autoscaling-min-processing-units=1000 --autoscaling-max-processing-units=5000 --autoscaling-high-priority-cpu-target=65

    The managed autoscaler is configured with the --autoscaling-* flags for minimum and maximum processing units and the CPU and storage targets. The other options use flags that do not exist.

  9. Question 10Design innovative, scalable, and highly available cloud database solutionsHard

    An AlloyDB cluster serves mixed transactional and reporting queries. Reporting queries degrade transactional latency on the primary instance. Which design change addresses this with the least application change?

    • A

      Move the reporting tables to Memorystore and query them using Redis commands

    • B

      Turn on the columnar engine on the primary and disable the write ahead log

    • C

      Increase the primary instance vCPU count until reporting no longer affects latency

    • D

      Add read pool instances and send reporting traffic to the read pool endpoint

    Show answer & explanation

    Correct answer: D — Add read pool instances and send reporting traffic to the read pool endpoint

    AlloyDB read pool instances scale read capacity horizontally and provide a separate endpoint, isolating reporting from the primary. Scaling the primary up is costly and does not isolate workloads. Memorystore is not a relational reporting store, and the write ahead log cannot be disabled, as it underpins durability.

Want more PCDE practice questions?

Take the free PCDE practice test: 30 questions per session with instant explanations. Unlock all 500 questions plus timed exam mode for $11.99 one-time.

Start the Free 30-Question Practice Test

Pricing

Choose Your Plan

Start free, upgrade when you're ready

Free

Good for exploring the platform

$0
  • 30 practice questions
  • Practice mode only
  • Progress tracking
  • No exam simulation mode
Start Free
POPULAR

PCDE Access

Best for the final weeks before your PCDE exam, after hands-on time with Google Cloud databases

$11.99one-time
  • 500 practice questions
  • Practice & Exam modes
  • Detailed explanations
  • Lifetime access
Get PCDE Access — $11.99

7-day money-back guarantee

BEST VALUE

Pro — All Certs

Planning multiple certifications

$9.99/month
  • ALL certifications included
  • Unlimited questions
  • New certs added free
  • Cancel anytime
View Pro Plans

Save 30% vs buying individually

Your PCDE Exam Won't Wait

Score 80% on two full CertSharp mock exams and you are calibrated to pass the live PCDE exam.

Don't risk $200 on the real exam without testing yourself first.

Start Free Practice Now

No credit card required · 30 free questions · Upgrade for $11.99 when ready

Official Certification Resources

Our practice questions align with official exam objectives. Study official materials first, then use CertSharp to test your readiness.

View Official PCDE Exam Details

FAQ

PCDE Frequently Asked Questions

Everything you need to know about Google Professional Cloud Database Engineer

What is the Google Professional Cloud Database Engineer certification?
The Google Professional Cloud Database Engineer (PCDE) validates your ability to design, create, manage, and troubleshoot Google Cloud databases used by applications to store and retrieve data. Cost: $200, 50-60 questions, 2-hour time limit.
How hard is the Google PCDE exam?
PCDE is a professional-level exam. It is scenario-driven: you choose between Cloud SQL, AlloyDB, Spanner, Bigtable, Firestore, and Memorystore, size and secure them, plan high availability, and pick the right migration approach. Knowing product names is not enough; you need to weigh cost, consistency, scale, and downtime tradeoffs.
Is Google PCDE worth it in 2026?
It is worth considering if your work centers on databases in Google Cloud, such as DBA, data platform, or migration roles. It is a focused credential that shows you can choose, run, and migrate databases on Google Cloud, rather than a broad architecture certification.
What experience does Google recommend for PCDE?
Google lists no prerequisites. It recommends 5+ years of overall database and IT experience, including 2 years of hands-on experience working with Google Cloud database solutions.
What topics are on the Google PCDE exam?
The exam guide has four sections: designing scalable and highly available database solutions (~32%), managing a solution that can span multiple database technologies (~25%), migrating data solutions (~23%), and deploying scalable and highly available databases (~20%). Topics include capacity planning, HA and DR, connectivity and IAM, monitoring, backup and PITR, replication, and migration tools.
How many questions are on the Google PCDE exam?
The exam has 50-60 multiple choice and multiple select questions with a 2-hour time limit. Google does not publish the passing score.
How long to prepare for Google PCDE?
It depends on your database background. Many candidates plan several weeks of study. Hands-on practice with Cloud SQL, AlloyDB, and Spanner, plus working through migration scenarios, helps most. Spend extra time on the design section, which carries the largest weight (~32%).
What is the difference between Google PCDE and Professional Data Engineer?
PCDE focuses on operational databases: choosing, sizing, securing, migrating, and running them. Professional Data Engineer (PDE) focuses on data pipelines, analytics, and machine learning workflows. If your role is database administration or migration, PCDE fits; if it is data processing and analytics, PDE fits.
How much does the Google Professional Cloud Database Engineer exam cost?
The Google Professional Cloud Database Engineer exam costs $200 USD, plus applicable taxes. It runs 2 hours and can be taken online-proctored from a remote location or at a testing center.
Is there a free Google PCDE practice test?
Yes. CertSharp offers a free Google PCDE practice test for the Google Professional Cloud Database Engineer exam: 30 questions per session with detailed explanations, no credit card required. The full 500-question Google PCDE bank, including timed exam mode, is $11.99 one-time with lifetime access.