Google Professional Cloud DevOps Engineer (GCP-PCDE) Certification Exam Sample Questions
We have prepared Google Professional Cloud DevOps Engineer (GCP-PCDE) certification sample questions to make you aware of actual exam properties. This sample question set provides you with information about the Google GCP-PCDE exam pattern, question formate, a difficulty level of questions and time required to answer each question. To get familiar with Google Cloud Platform - Professional Cloud DevOps Engineer (GCP-PCDE) exam, we suggest you try our Sample Google GCP-PCDE Certification Practice Exam in simulated Google certification exam environment.
To test your knowledge and understanding of concepts with real-time scenario based questions, we strongly recommend you to prepare and practice with Premium Google GCP-PCDE Certification Practice Exam. The premium certification practice exam helps you identify topics in which you are well prepared and topics in which you may need further training to achieving great score in actual Google Cloud Platform - Professional Cloud DevOps Engineer (GCP-PCDE) exam.
Google GCP-PCDE Sample Questions:
01. Terraform changes to production are reviewed as pull requests, but reviewers judge them from the diff alone and have twice missed that a change would destroy and recreate a database instance. The team wants reviewers to see the consequence rather than only the code.
What should the pipeline add?
a) An apply step on every pull request against a temporary copy of production, torn down once the reviewer has looked at the resulting environment
b) A required second reviewer on any pull request that touches a file under the database module directory
c) A lifecycle prevention block on the database instance, so an apply that would destroy it fails when it runs
d) A plan step on every pull request that posts the plan to the request, with a policy check that fails on a protected resource being destroyed
02. A GKE node pool of twelve nodes runs a service whose pod disruption budget permits one unavailable replica. Upgrades currently take several hours. The team wants them faster, without breaching the budget and without exceeding the two nodes of CPU quota headroom the project has.
Which node upgrade setting fits?
a) Surge upgrade with a maximum surge of zero and a maximum unavailable of two
b) Surge upgrade with a maximum surge of six and a maximum unavailable of zero
c) Surge upgrade with a maximum surge of two and a maximum unavailable of zero
d) Blue-green upgrade with the full node pool duplicated before any workload is moved
03. An on-call rotation spends roughly eight hours a week restarting a stuck consumer, clearing its queue and filing the same ticket each time. The team has capacity to automate one thing this quarter.
Which characteristic identifies this as the right candidate?
a) It is manual, repetitive, automatable and grows in proportion to the service
b) It is unplanned, undocumented, unowned and outside the team's published support scope
c) It is urgent, customer-visible, unpredictable and handled by the most experienced engineer available
d) It is frequent, low-risk, well understood and already covered by an existing runbook
04. Several engineers now run Terraform against the same production configuration from their workstations. Two applies have already collided and left resources half-created.
Which change prevents this?
a) Commit the state file to the shared Git repository, which makes every change reviewable before it lands
b) Move the state to a Cloud Storage backend, which locks the state object for the duration of an apply
c) Split the configuration into one workspace per engineer, which keeps each apply in its own state
d) Schedule applies through a nightly job, which removes the chance of two engineers running at once
05. Every new delivery team needs the same set of projects created the same way: the right folder, the billing account, a baseline set of enabled APIs, a log sink and the standard IAM bindings. A platform engineer does this by hand today, and the details have drifted between teams.
What should the team adopt?
a) A reusable Terraform module that provisions the whole project set, invoked with per-team input variables
b) A gcloud script that creates the projects, with the remaining baseline settings applied by each team afterwards
c) A documented runbook in the internal wiki that the platform engineer follows for each new team
d) An organization policy that enforces the baseline settings on any project once it has been created
06. Access for a delivery team is granted today by binding each engineer's user account directly to roles in the team's projects. Onboarding and offboarding take a day of manual work, and quarterly access reviews keep finding stale grants.
What should the platform team change?
a) Bind roles to individual accounts as now, adding an IAM condition that expires each binding after ninety days
b) Bind roles to Google groups that mirror the team's job functions, managing membership in Cloud Identity
c) Bind roles at the folder level to the same individual accounts, so one change covers every project underneath
d) Bind roles to a single shared service account that engineers impersonate, keeping the impersonation list per project
07. Two services owned by different teams share one project. The teams argue over quota, and an IAM change made for one service keeps granting the other more access than it needs. Someone proposes splitting them into separate projects.
What does that split actually provide?
a) Separate quota pools, separate IAM policy surfaces and separate audit scopes
b) Automatic network isolation, separate firewall defaults and blocked cross-project traffic
c) Lower inter-service latency, higher available bandwidth and fewer network hops
d) A smaller combined bill through per-project discounts and committed-use pricing
08. One Terraform state currently holds development, staging and production for a service. A recent production apply also recreated a development resource that had been deleted by hand, and the run took twenty minutes.
What should the team do?
a) Keep one state and target individual resources with the resource-targeting flag on every production apply
b) Keep one state and raise the parallelism setting to shorten the run, with a manual review of every plan
c) Split the configuration into one state per environment, each with its own backend prefix and pipeline
d) Split the configuration into one state per resource type, so a change touches only the relevant state
09. A user request passes through an API gateway, an authentication service, an orders service and a datastore. Latency at the ninety-ninth percentile has doubled over a week. Each service's own latency dashboard looks unchanged.
Which signal identifies where the extra time is going?
a) A profile from each service, showing which functions consume the most processor time
b) An uptime check against each service's health endpoint, sampled from several regions
c) A log-based metric on each service's request duration field, charted side by side
d) A distributed trace of slow requests in Cloud Trace, showing span durations across the four hops
10. Cloud Logging ingestion cost has tripled. Investigation shows that debug-level entries from one service account for most of the volume, and nobody has queried them in six months. Audit logs must be retained.
What should the team do?
a) Lower the retention on the default log bucket to thirty days, leaving ingestion as it is
b) Route that service's logs to a Cloud Storage bucket through a sink, while also keeping them in Cloud Logging
c) Add an exclusion filter on the Log Router matching that service's debug entries, leaving other logs untouched
d) Switch off the service's debug logging in its application configuration and redeploy it
Answers:
|
Question: 01 Answer: d |
Question: 02 Answer: c |
Question: 03 Answer: a |
Question: 04 Answer: b |
Question: 05 Answer: a |
|
Question: 06 Answer: b |
Question: 07 Answer: a |
Question: 08 Answer: c |
Question: 09 Answer: d |
Question: 10 Answer: c |
Note: Please update us by writing an email on feedback@vmexam.com for any error in Google Cloud Platform - Professional Cloud DevOps Engineer (GCP-PCDE) certification exam sample questions
