Road to Google Cloud Architect Certification Flashcards
1
Q
Building for Builders LLC manufactures equipment used in residential and commercial
building. Each of its 500,000 pieces of equipment in use around the globe has IoT devices
collecting data about the state of equipment. The IoT data is streamed from each device
every 10 seconds. On average, 10 KB of data is sent in each message. The data will be used
for predictive maintenance and product development. The company would like to use a
managed database in Google Cloud. What would you recommend?
A. Apache Cassandra
B. Cloud Bigtable
C. BigQuery
D. CloudSQL
A
B. Option B is correct. Bigtable is the best option for streaming IoT data, since it supports
low-latency writes and is designed to scale to support petabytes of data. Option A is incorrect because Apache Cassandra is not a managed database in GCP. Option C is incorrect
because BigQuery is an analytics database. While it is a good option for analyzing the data,
Bigtable is a better option for ingesting the data. Option D is incorrect. CloudSQL is a
managed relational database. The use case does not require a relational database, and Bigtable’s scalability is a better fit with the requirements.
2
Q
You have developed a web application that is becoming widely used. The frontend runs in
Google App Engine and scales automatically. The backend runs on Compute Engine in a
managed instance group. You have set the maximum number of instances in the backend
managed instance group to five. You do not want to increase the maximum size of the managed instance group or change the VM instance type, but there are times the frontend sends
more data than the backend can keep up with and data is lost. What can you do to prevent
the loss of data?
A. Use an unmanaged instance group B. Store ingested data in Cloud Storage C. Have the frontend write data to a Cloud Pub/Sub topic, and have the backend read from that topic D. Store ingested data in BigQuery
A
C. The correct answer is C. A Cloud Pub/Sub topic would decouple the frontend and
backend, provide a managed and scalable message queue, and store ingested data until the
backend can process it. Option A is incorrect. Switching to an unmanaged instance group
will mean that the instance group cannot autoscale. Option B is incorrect. You could store
ingested data in Cloud Storage, but it would not be as performant as the Cloud Pub/Sub
solution. Option D is incorrect because BigQuery is an analytics database and not designed
for this use case.
3
Q
You are setting up a cloud project and want to assign members of your team different permissions. What GCP service would you use to do that?
A. Cloud Identity
B. Identity and Access Management (IAM)
C. Cloud Authorizations
D. LDAP
A
B. The correct answer is B. IAM is used to manage roles and permissions. Option A is
incorrect. Cloud Identity is a service for creating and managing identities. Option C is
incorrect. There is no GCP service with that name at this time. Option D is incorrect.
LDAP is not a GCP service.
4
Q
You would like to run a custom container in a managed Google Cloud Service. What are
your two options?
A. App Engine Standard and Kubernetes Engine
B. App Engine Flexible and Kubernetes Engine
C. Compute Engine and Kubernetes Engine
D. Cloud Functions and App Engine Flexible
A
B. The correct answer is B. You can run custom containers in App Engine Flexible and
Kubernetes Engine. Option A is incorrect because App Engine Standard does not support
custom containers. Option C is incorrect because Compute Engine is not a managed service. Option D is incorrect because Cloud Functions does not support custom containers.
5
Q
PhotosForYouToday prints photographs and ships them to customers. The frontend application uploads photos to Cloud Storage. Currently, the backend runs a cron job that checks
Cloud Storage buckets every 10 minutes for new photos. The product manager would like
to process the photos as soon as they are uploaded. What would you use to cause processing to start when a photo file is saved to Cloud Storage?
A Cloud Function
B. An App Engine Flexible application
C. A Kubernetes pod
D. A cron job that checks the bucket more frequently
A
A. The correct answer is A. A Cloud Function can respond to a create file event in Cloud
Storage and start processing when the file is created. Option B is incorrect because an
App Engine Flexible application cannot respond to a Cloud Storage write event. Option
C is incorrect. Kubernetes pods are the smallest compute unit in Kubernetes and are not
designed to respond to Cloud Storage events. Option D is incorrect because it does not
guarantee that photos will be processed as soon as they are created.
6
Q
The chief financial officer of your company believes that you are spending too much money
to run an on-premises data warehouse and wants to migrate to a managed cloud solution.
What GCP service would you recommend for implementing a new data warehouse in GCP?
A. Compute Engine
B. BigQuery
C. Cloud Dataproc
D. Cloud Bigtable
A
B. The correct answer is B. BigQuery is a managed analytics database designed to support
data warehouses and similar use cases. Option A is incorrect. Compute Engine is not a
managed service. Option C is incorrect. Cloud Dataproc is a managed Hadoop and Spark
service. Option D is incorrect. Bigtable is a NoSQL database well suited for large-volume,
low-latency writes and limited ranges of queries. It is not suitable for the kind of ad hoc
querying commonly done with data warehouses.
7
Q
A government regulation requires you to keep certain financial data for seven years. You
are not likely to ever retrieve the data, and you are only keeping it to be in compliance.
There are approximately 500 GB of financial data for each year that you are required to
save. What is the most cost-effective way to store this data?
A. Cloud Storage multiregional storage
B. Cloud Storage Nearline storage
C. Cloud Storage Coldline storage
D. Cloud Storage persistent disk storage
A
C. The correct answer is C. Cloud Storage Coldline is the lowest-cost option, and it is
designed for data that is accessed less than once per year. Option A and Option B are incorrect because they cost more than Coldline storage. Option D is incorrect because there is no
such service.
8
Q
Global Games Enterprises Inc. is expanding from North America to Europe. Some of the
games offered by the company collect personal information. With what additional regulation will the company need to comply when it expands into the European market?
A. HIPAA
B. PCI-DS
C. GDPR
D. SOX
A
C. The correct answer is C. The GDPR is a European Union directive protecting the personal information of EU citizens. Option A is incorrect. HIPAA is a U.S. healthcare regulation. Option B is incorrect. PCI-DS is a payment card data security regulation; if Global
Games Enterprises Inc. is accepting payment cards in North America, it is already subject
to that regulation. Option D is a U.S. regulation on some publicly traded companies; the
company may be subject to that regulation already, and expanding to Europe will not
change its status.
9
Q
Your team is developing a Tier 1 application for your company. The application will depend
on a PostgreSQL database. Team members do not have much experience with PostgreSQL
and want to implement the database in a way that minimizes their administrative responsibilities for the database. What managed service would you recommend?
A. Cloud SQL
B. Cloud Dataproc
C. Cloud Bigtable
D. Cloud PostgreSQL
A
A. The correct answer is A. Cloud SQL is a managed database service that supports PostgreSQL. Option B is incorrect. Cloud Dataproc is a managed Hadoop and Spark service.
Option C is incorrect. Cloud Bigtable is a NoSQL database. Option D is incorrect. There is
no service called Cloud PostgreSQL in GCP at this time
10
Q
What is a service-level indicator?
A. A metric collected to indicate how well a service-level objective is being met
B. A type of log
C. A type of notification sent to a sysadmin when an alert is triggered
D. A visualization displayed when a VM instance is down
A
A. The correct answer is A. A service-level indicator is a metric used to measure how well
a service is meeting its objectives. Options B and C are incorrect. It is not a type of log or a
type of notification. Option D is incorrect. A service-level indicator is not a visualization,
although the same metrics may be used to drive the display of a visualization.
11
Q
Developers at MakeYouFashionable have adopted agile development methodologies. Which
tool might they use to support CI/CD?
A. Google Docs
B. Jenkins
C. Apache Cassandra
D. Clojure
A
B. The correct answer is B. Jenkins is a popular CI/CD tool. Option A is incorrect. Google
Docs is a collaboration tool for creating and sharing documents. Option C is incorrect.
Cassandra is a NoSQL database. Option D is incorrect. Clojure is a Lisp-like programming
language that runs on the Java virtual machine (JVM).
12
Q
You have a backlog of audio files that need to be processed using a custom application.
The files are stored in Cloud Storage. If the files were processed continuously on three
n1-standard-4 instances, the job could complete in two days. You have 30 days to deliver
the processed files, after which they will be sent to a client and deleted from your systems.
You would like to minimize the cost of processing. What might you do to help keep costs
down?
A. Store the files in coldline storage
B. Store the processed files in multiregional storage
C. Store the processed files in Cloud CDN
D. Use preemptible VMs
A
D. The correct answer is D. Use preemptible VMs, which cost significantly less than
standard VMs. Option A is incorrect. Coldline storage is not appropriate for files that are
actively used. Option B is incorrect. Storing files in multiregional storage will cost more
than regional storage, and there is no indication from the requirements that they should be
stored multiregionally. Option C is incorrect. There is no indication that the processed files
need to be distributed to a global user base.
13
Q
You have joined a startup selling supplies to visual artists. One element of the company’s
strategy is to foster a social network of artists and art buyers. The company will provide
e-commerce services for artists and earn revenue by charging a fee for each transaction.
You have been asked to collect more detailed business requirements. What might you
expect as an additional business requirement?
A. The ability to ingest streaming data
B. A recommendation system to match buyers to artists
C. Compliance with SOX regulations
D. Natural language processing of large volumes of text
A
B. The correct answer is B. This is an e-commerce site matching sellers and buyers, so a
system that recommends artists to buyers can help increase sales. Option A is incorrect.
There is no indication of any need for streaming data. Option C is incorrect. This is a
startup, and it is not likely subject to SOX regulations. Option D is incorrect. There is no
indication of a need to process large volumes of text
14
Q
You work for a manufacturer of specialty die cast parts for the aerospace industry. The
company has built a reputation as the leader in high-quality, specialty die cast parts, but
recently the number of parts returned for poor quality is increasing. Detailed data about the
manufacturing process is collected throughout every stage of manufacturing. To date, the
data has been collected and stored but not analyzed. There are a total of 20 TB of data. The
company has a team of analysts familiar with spreadsheets and SQL. What service might
you recommend for conducting preliminary analysis of the data?
A. Compute Engine
B. Kubernetes Engine
C. BigQuery
D. Cloud Functions
A
C. The correct answer is C. BigQuery is an analytics database that supports SQL. Options
A and B are incorrect because, although they could be used to run analytics applications,
such as Apache Hadoop or Apache Spark, it would require more administrative overhead.
Also, the team members working on this are analysts, but there is no indication that they
have the skills or desire to manage analytics platforms. Option D is incorrect. Cloud Functions is for running short programs in response to events in GCP.
15
Q
A client of yours wants to run an application in a highly secure environment. They want to
use instances that will only run boot components verified by digital signatures. What would
you recommend they use in Google Cloud?
A. Preemptible VMs
B. Managed instance groups
C. Cloud Functions
D. Shielded VMs
A
The correct answer is D.
Shielded VMs include secure boot, which only runs digitally verified boot components. Option A is incorrect. Preemptible VMs are interruptible
instances, but they cost less than standard VMs. Option B is incorrect. Managed instance
groups are sets of identical VMs that are managed as a single entity. Option C is incorrect.
Cloud Functions is a PaaS for running programs in response to events in GCP.
16
Q
You have installed the Google Cloud SDK. You would now like to work on transferring
files to Cloud Storage. What command-line utility would you use?
A. bq
B. gsutil
C. cbt
D. gcloud
A
B. The correct answer is B. gsutil is the command-line utility for working with Cloud
Storage. Option A is incorrect. bq is the command-line utility for working with BigQuery.
Option C is incorrect. cbt is the command-line utility for working with Cloud Bigtable.
Option D is incorrect. gcloud is used to work with most GCP services but not Cloud
Storage.
17
Q
Kubernetes pods sometimes need access to persistent storage. Pods are ephemeral—they
may shut down for reasons not in control of the application running in the pod. What
mechanism does Kubernetes use to decouple pods from persistent storage?
A. PersistentVolumes
B. Deployments
C. ReplicaSets
D. Ingress
A
A. The correct answer is A. PersistentVolumes is Kubernetes’ way of representing storage
allocated or provisioned for use by a pod. Option B is incorrect. Deployments are a type
of controller consisting of pods running the same version of an application. Option C
is incorrect. A ReplicaSet is a controller that manages the number of pods running in a
deployment. Option D is incorrect. An Ingress is an object that controls external access to
services running in a Kubernetes cluster.
18
Q
An application that you support has been missing service-level objectives, especially around
database query response times. You have reviewed monitoring data and determined that
a large number of database read operations is putting unexpected load on the system. The
database uses MySQL, and it is running in Compute Engine. You have tuned SQL queries,
and the performance is still not meeting objectives. Of the following options, which would
you try next?
A. Migrate to a NoSQL database.
B. Move the database to Cloud SQL.
C. Use Cloud Memorystore to cache data read from the database to reduce the number of
reads on the database.
D. Move some of the data out of the database to Cloud Storage.
A
C. The correct answer is C. Use Cloud Memorystore to reduce the number of reads against
the database. Option A is incorrect. The application is designed to work with a relational
database, and there is no indication that a NoSQL database is a better option overall.
Option B is incorrect. Simply moving the database to a managed service will not change the
number of read operations, which is the cause of the poor performance. Option D is incorrect. Moving data to Cloud Storage will not reduce the number of reads
19
Q
You are running a complicated stream processing operation using Apache Beam. You want
to start using a managed service. What GCP service would you use?
A. Cloud Dataprep
B. Cloud Dataproc
C. Cloud Dataflow
D. Cloud Identity
A
C. The correct answer is C. Cloud Dataflow is an implementation of the Apache Beam
stream processing framework. Cloud Dataflow is a fully managed service. Option A is
incorrect. Cloud Dataprep is used to prepare data for analysis. Option B is incorrect. Cloud
Dataproc is a managed Hadoop and Spark service. Option D is incorrect. Cloud Identity is
an authentication service.
20
Q
Your team has had a number of incidents in which Tier 1 and Tier 2 services were down for
more than 1 hour. After conducting a few retrospective analyses of the incidents, you have
determined that you could identify the causes of incidents faster if you had a centralized log
repository. What GCP service could you use for this?
A. Stackdriver Logging
B. Cloud Logging
C. Cloud SQL
D. Cloud Bigtable
A
A. The correct answer is A. Stackdriver Logging is a centralized logging service. Option
B is incorrect. There is no such service at this time. Option C and Option D are incorrect
because those are databases and not specifically designed to support the logging of the use
case described.
21
Q
A Global 2000 company has hired you as a consultant to help architect a new logistics system. The system will track the location of parts as they are shipped between company facilities in Europe, Africa, South America, and Australia. Anytime a user queries the database,
they must receive accurate and up-to-date information; specifically, the database must support strong consistency. Users from any facility may query the database using SQL. What
GCP service would you recommend?
A. Cloud SQL
B. BigQuery
C. Cloud Spanner
D. Cloud Dataflow
A
C. The correct answer is C. Cloud Spanner is a globally scalable, strongly consistent relational database that can be queried using SQL. Option A is incorrect because it will not
scale to the global scale as Cloud Spanner will. Option B is incorrect. The requirements
describe an application that will likely have frequent updates and transactions. BigQuery
is designed for analytics and data warehousing. Option D is incorrect. Cloud Dataflow is a
stream and batch processing service.
22
Q
A database architect for a game developer has determined that a NoSQL document
database is the best option for storing players’ possessions. What GCP service would you
recommend?
A. Cloud Datastore
B. Cloud Storage
C. Cloud Dataproc
D. Cloud Bigtable
A
A. The correct answer is A. Cloud Datastore is a managed document NoSQL database
in GCP. Option B is incorrect. Cloud Storage is an object storage system, not a document
NoSQL database. Option C is incorrect. Cloud Dataproc is a managed Hadoop and Spark
service. Option D is incorrect. Cloud Bigtable is a wide-column NoSQL database, not a
document database.
23
Q
A major news agency is seeing increasing readership across the globe. The CTO is concerned that long page-load times will decrease readership. What might the news agency try
to reduce the page-load time of readers around the globe?
A. Regional Cloud Storage
B. Cloud CDN
C. Fewer firewall rules
D. Virtual private network
A
B. The correct answer is B. Cloud CDN is GCP’s content delivery network, which distributes static content globally. Option A is incorrect. Reading from regional storage can still
have long latencies for readers outside of the region. Option C is incorrect. Firewall rules do
not impact latency in any discernible way. Option D is incorrect. VPNs are used to link onpremises networks to Google Cloud
24
Q
What networking mechanism allows different VPC networks to communicate using private
IP address space, as defined in RFC 1918?
A. ReplicaSets
B. Custom subnets
C. VPC network peering
D. Firewall rules
A
C. The correct answer is C. VPC peering allows different VPCs to communicate using
private networks. Option A is incorrect. ReplicaSets are used in Kubernetes; they are not
related to VPCs. Option B is incorrect. Custom subnets define network address ranges for
regions. Option D is incorrect. Firewall rules control the flow of network traffic.
25
You have been tasked with setting up disaster recovery infrastructure in the cloud that will
be used if the on-premises data center is not available. What network topology would you
use for a disaster recovery environment?
A. Meshed topology
B. Mirrored topology
C. Gated egress topology
D. Gated ingress topology
B. The correct answer is B. With a mirrored topology, the public cloud and private onpremise environments mirror each other. Option A is incorrect. In a mesh topology, all
systems in the cloud and private networks can communicate with each other. Option C is
incorrect. In a gated egress topology, on-premises service APIs are made available to applications running in the cloud without exposing them to the public Internet. Option D is
incorrect. In a gated ingress topology, cloud service APIs are made available to applications
running on-premises without exposing them to the public Internet.
26
For this question, refer to the TerramEarth case study. https://cloud.google.com/certification/guides/cloud-architect/casestudy-terramearth-rev2
Because you do not know every possible future use for the data TerramEarth collects, you have decided to build a system that captures and stores all raw data in case you need it later. How can you most cost-effectively accomplish this goal?
A. Have the vehicles in the field stream the data directly into BigQuery.
B. Have the vehicles in the field pass the data to Cloud Pub/Sub and dump it into a Cloud Dataproc cluster that stores data in Apache Hadoop Distributed File System (HDFS) on persistent disks.
C. Have the vehicles in the field continue to dump data via FTP, adjust the existing Linux machines, and use a collector to upload them into Cloud Dataproc HDFS for storage.
D. Have the vehicles in the field continue to dump data via FTP, and adjust the existing Linux machines to immediately upload it to Cloud Storage with gsutil.
D. Have the vehicles in the field continue to dump data via FTP, and adjust the existing Linux machines to immediately upload it to Cloud Storage with gsutil.
Commentaire
A is not correct because TerramEarth has cellular service for 200,000 vehicles, and each vehicle sends at least one row (120 fields) per second. This exceeds BigQuery's maximum rows per second per project quota. Additionally, there are 20 million total vehicles, most of which perform uploads when connected by a maintenance port, which drastically exceeds the streaming project quota further.
B is not correct because although Cloud Pub/Sub is a fine choice for this application, Cloud Dataproc is probably not. The question posed asks us to optimize for cost. Because Cloud Dataproc is optimized for ephemeral, job-scoped clusters, a long-running cluster with large amounts of HDFS storage could be very expensive to build and maintain when compared to managed and specialized storage solutions like Cloud Storage.
C is not correct because the question asks us to optimize for cost, and because Cloud Dataproc is optimized for ephemeral, job-scoped clusters, a long-running cluster with large amounts of HDFS storage could be very expensive to build and maintain when compared to managed and specialized storage solutions like Cloud Storage.
D is correct because several load-balanced Compute Engine VMs would suffice to ingest 9 TB per day, and Cloud Storage is the cheapest per-byte storage offered by Google. Depending on the format, the data could be available via BigQuery immediately, or shortly after running through an ETL job. Thus, this solution meets business and technical requirements while optimizing for cost.
https: //cloud.google.com/blog/products/data-analytics/10-tips-for-building-long-running-clusters-using-cloud-dataproc
https: //cloud.google.com/blog/products/data-analytics/10-tips-for-building-long-running-clusters-using-cloud-dataproc
https: //cloud.google.com/bigquery/quotas#streaming_inserts
27
For this question, refer to the TerramEarth case study. https://cloud.google.com/certification/guides/cloud-architect/casestudy-terramearth-rev2
Today, TerramEarth maintenance workers receive interactive performance graphs for the last 24 hours (86,400 events) by plugging their maintenance tablets into the vehicle. The support group wants support technicians to view this data remotely to help troubleshoot problems. You want to minimize the latency of graph loads. How should you provide this functionality?
A. Execute queries against data stored in a Cloud SQL.
B. Execute queries against data indexed by vehicle_id.timestamp in Cloud Bigtable.
C. Execute queries against data stored on daily partitioned BigQuery tables.
D. Execute queries against BigQuery with data stored in Cloud Storage via BigQuery federation.
A is not correct because Cloud SQL provides relational database services that are well suited to OLTP workloads, but not storage and low-latency retrieval of time-series data.
B is correct because Cloud Bigtable is optimized for time-series data. It is cost-efficient, highly available, and low-latency. It scales well. Best of all, it is a managed service that does not require significant operations work to keep running.
C is not correct because BigQuery is fast for wide-range queries, but it is not as well optimized for narrow-range queries as Cloud Bigtable is. Latency will be an order of magnitude shorter with Cloud Bigtable for this use.
D is not correct because the objective is to minimize latency, and although BigQuery federation offers tremendous flexibility, it doesn't perform as well as native BigQuery storage, and will have longer latency than Cloud Bigtable for narrow-range queries.
https: //cloud.google.com/bigquery/external-data-sources
https: //cloud.google.com/bigtable/docs/schema-design-time-series#time-series-cloud-bigtable
28
For this question, refer to the TerramEarth case study. https://cloud.google.com/certification/guides/cloud-architect/casestudy-terramearth-rev2
Your agricultural division is experimenting with fully autonomous vehicles. You want your architecture to promote strong security during vehicle operation. Which two architecture characteristics should you consider? (choose two)
A. Use multiple connectivity subsystems for redundancy.
B. Require IPv6 for connectivity to ensure a secure address space.
C. Enclose the vehicle’s drive electronics in a Faraday cage to isolate chips.
D. Use a functional programming language to isolate code execution cycles.
E. Treat every microservice call between modules on the vehicle as untrusted.
F. Use a Trusted Platform Module (TPM) and verify firmware and binaries on boot.
Bonne réponse
E. Treat every microservice call between modules on the vehicle as untrusted.
F. Use a Trusted Platform Module (TPM) and verify firmware and binaries on boot.
Commentaire
A is not correct because this improves system durability, but it doesn't have any impact on the security during vehicle operation.
B is not correct because IPv6 doesn't have any impact on the security during vehicle operation, although it improves system scalability and simplicity.
C is not correct because it doesn't have any impact on the security during vehicle operation, although it improves system durability.
D is not correct because merely using a functional programming language doesn't guarantee a more secure level of execution isolation. Any impact on security from this decision would be incidental at best.
E is correct because this improves system security by making it more resistant to hacking, especially through man-in-the-middle attacks between modules.
F is correct because this improves system security by making it more resistant to hacking, especially rootkits or other kinds of corruption by malicious actors.
https://en.wikipedia.org/wiki/Trusted_Platform_Module
29
For this question, refer to the TerramEarth case study. https://cloud.google.com/certification/guides/cloud-architect/casestudy-terramearth-rev2
Which of TerramEarth’s legacy enterprise processes will experience significant change as a result of increased Google Cloud Platform adoption?
A. OpEx/CapEx allocation, LAN change management, capacity planning
B. Capacity planning, TCO calculations, OpEx/CapEx allocation
C. Capacity planning, utilization measurement, data center expansion
D. Data center expansion,TCO calculations, utilization measurement
Commentaire
A is not correct because LAN change management processes don't need to change significantly. TerramEarth can easily peer their on-premises LAN with their Google Cloud Platform VPCs, and as devices and subnets move to the cloud, the LAN team's implementation will change, but the change management process doesn't have to.
B is correct because all of these tasks are big changes when moving to the cloud. Capacity planning for cloud is different than for on-premises data centers; TCO calculations are adjusted because TerramEarth is using services, not leasing/buying servers; OpEx/CapEx allocation is adjusted as services are consumed vs. using capital expenditures.
C is not correct because measuring utilization can be done in the same way, often with the same tools (along with some new ones). Data center expansion is not a concern for cloud customers; it is part of the undifferentiated heavy lifting that is taken care of by the cloud provider.
D is not correct because data center expansion is not a concern for cloud customers; it is part of the undifferentiated heavy lifting that is taken care of by the cloud provider. Measuring utilization can be done in the same way, often with the same tools (along with some new ones).
https://assets.kpmg/content/dam/kpmg/pdf/2015/11/cloud-economics.pdf
30
For this question, refer to the TerramEarth case study. https://cloud.google.com/certification/guides/cloud-architect/casestudy-terramearth-rev2
You analyzed TerramEarth’s business requirement to reduce downtime and found that they can achieve a majority of time saving by reducing customers’ wait time for parts. You decided to focus on reduction of the 3 weeks’ aggregate reporting time. Which modifications to the company’s processes should you recommend?
A. Migrate from CSV to binary format, migrate from FTP to SFTP transport, and develop machine learning analysis of metrics.
B. Migrate from FTP to streaming transport, migrate from CSV to binary format, and develop machine learning analysis of metrics.
C. Increase fleet cellular connectivity to 80%, migrate from FTP to streaming transport, and develop machine learning analysis of metrics.
D. Migrate from FTP to SFTP transport, develop machine learning analysis of metrics, and increase dealer local inventory by a fixed factor.
Bonne réponse
C. Increase fleet cellular connectivity to 80%, migrate from FTP to streaming transport, and develop machine learning analysis of metrics.
Commentaire
A is not correct because machine learning analysis is a good means toward the end of reducing downtime, but shuffling formats and transport doesn't directly help at all.
B is not correct because machine learning analysis is a good means toward the end of reducing downtime, and moving to streaming can improve the freshness of the information in that analysis, but changing the format doesn't directly help at all.
C is correct because using cellular connectivity will greatly improve the freshness of data used for analysis from where it is now, collected when the machines are in for maintenance. Streaming transport instead of periodic FTP will tighten the feedback loop even more. Machine learning is ideal for predictive maintenance workloads.
D is not correct because machine learning analysis is a good means toward the end of reducing downtime, but the rest of these changes don't directly help at all.
31
Your company wants to deploy several microservices to help their system handle elastic loads. Each microservice uses a different version of software libraries. You want to enable their developers to keep their development environment in sync with the various production services. Which technology should you choose?
A. RPM/DEB
B. Containers
C. Chef/Puppet
D. Virtual machines
A is not correct because although OS packages are a convenient way to distribute and deploy libraries, they don't directly help with synchronizing. Even with a common repository, the development environments will probably deviate from production.
B is correct because using containers for development, test, and production deployments abstracts away system OS environments, so that a single host OS image can be used for all environments. Changes that are made during development are captured using a copy on-write filesystem, and teams can easily publish new versions of the microservices in a repository.
C is not correct because although infrastructure configuration as code can help unify production and test environments, it is very difficult to make all changes during development this way.
D is not correct because virtual machines run their own OS, which will eventually deviate in each environment, just as now.
32
Your company wants to track whether someone is present in a meeting room reserved for a scheduled meeting. There are 1000 meeting rooms across 5 offices on 3 continents. Each room is equipped with a motion sensor that reports its status every second. You want to support the data ingestion needs of this sensor network. The receiving infrastructure needs to account for the possibility that the devices may have inconsistent connectivity. Which solution should you design?
A. Have each device create a persistent connection to a Compute Engine instance and write messages to a custom application.
B. Have devices poll for connectivity to Cloud SQL and insert the latest messages on a regular interval to a device specific table.
C. Have devices poll for connectivity to Cloud Pub/Sub and publish the latest messages on a regular interval to a shared topic for all devices.
D. Have devices create a persistent connection to an App Engine application fronted by Cloud Endpoints, which ingest messages and write them to Cloud Datastore.
A is not correct because having a persistent connection does not handle the case where the device is disconnected.
B is not correct because Cloud SQL is a regional, relational database and not the best fit for sensor data. Additionally, the frequency of the writes has the potential to exceed the supported number of concurrent connections.
C is correct because Cloud Pub/Sub can handle the frequency of this data, and consumers of the data can pull from the shared topic for further processing.
D is not correct because having a persistent connection does not handle the case where the device is disconnected.
https: //cloud.google.com/sql/
https: //cloud.google.com/pubsub/
33
Your company wants to try out the cloud with low risk. They want to archive approximately 100 TB of their log data to the cloud and test the serverless analytics features available to them there, while also retaining that data as a long-term disaster recovery backup. Which two steps should they take? (choose two)
```
A. Load logs into BigQuery.
B. Load logs into Cloud SQL.
C. Import logs into Cloud Logging.
D. Insert logs into Cloud Bigtable.
E. Upload log files into Cloud Storage.
```
A is correct because BigQuery is a serverless warehouse for analytics and supports the volume and analytics requirement.
B is not correct because Cloud SQL does not support the expected 100 TB. Additionally, Cloud SQL is a relational database and not the best fit for time-series log data formats.
C is not correct because Cloud Logging is optimized for monitoring, error reporting, and debugging instead of analytics queries.
D is not correct because Cloud Bigtable is optimized for read-write latency and analytics throughput, not analytics querying and reporting.
E is correct because Cloud Storage provides the Coldline and Archive storage classes to support long-term storage with infrequent access, which would support the long-term disaster recovery backup requirement.
https: //cloud.google.com/storage/docs/storage-classes#coldline
https: //cloud.google.com/bigtable/
https: //cloud.google.com/products/operations
https: //cloud.google.com/sql/
https: //cloud.google.com/bigquery/
34
You set up an autoscaling managed instance group to serve web traffic for an upcoming launch. After configuring the instance group as a backend service to an HTTP(S) load balancer, you notice that virtual machine (VM) instances are being terminated and re-launched every minute. The instances do not have a public IP address. You have verified that the appropriate web response is coming from each instance using the curl command. You want to ensure that the backend is configured correctly. What should you do?
A. Ensure that a firewall rule exists to allow source traffic on HTTP/HTTPS to reach the load balancer.
B. Assign a public IP to each instance, and configure a firewall rule to allow the load balancer to reach the instance public IP.
C. Ensure that a firewall rule exists to allow load balancer health checks to reach the instances in the instance group.
D. Create a tag on each instance with the name of the load balancer. Configure a firewall rule with the name of the load balancer as the source and the instance tag as the destination.
A is not correct because the issue to resolve is the VMs being terminated, not access to the load balancer.
B is not correct because this introduces a security vulnerability without addressing the primary concern of the VM termination.
C is correct because health check failures lead to a VM being marked unhealthy and can result in termination if the health check continues to fail. Because you have already verified that the instances are functioning properly, the next step would be to determine why the health check is continuously failing.
D is not correct because the source of the firewall rule that allows load balancer and health check access to instances is defined IP ranges, and not a named load balancer. Tagging the instances for the purpose of firewall rules is appropriate but would probably be a descriptor of the application, and not the load balancer.
https: //cloud.google.com/load-balancing/docs/https/
https: //cloud.google.com/load-balancing/docs/health-check-concepts
35
You are designing a large distributed application with 30 microservices. Each of your distributed microservices needs to connect to a database backend. You want to store the credentials securely. Where should you store the credentials?
A. In the source code
B. In an environment variable
C. In a secret management system
D. In a config file that has restricted access through ACLs
A is not correct because storing credentials in source code and source control is discoverable, in plain text, by anyone with access to the source code. This also introduces the requirement to update code and do a deployment each time the credentials are rotated.
B is not correct because consistently populating environment variables would require the credentials to be available, in plain text, when the session is started.
C is correct because a secret management system such as Secret Manager is a secure and convenient storage system for API keys, passwords, certificates, and other sensitive data. Secret Manager provides a central place and single source of truth to manage, access, and audit secrets across Google Cloud.
D is not correct because instead of managing access to the config file and updating manually as keys are rotated, it would be better to leverage a key management system. Additionally, there is increased risk if the config file contains the credentials in plain text.
https: //cloud.google.com/kubernetes-engine/docs/concepts/secret
https: //cloud.google.com/secret-manager
36
Mountkirk Games wants to set up a real-time analytics platform for their new game. The new platform must meet their technical requirements. Which combination of Google technologies will meet all of their requirements?
A. Kubernetes Engine, Cloud Pub/Sub, and Cloud SQL
B. Cloud Dataflow, Cloud Storage, Cloud Pub/Sub, and BigQuery
C. Cloud SQL, Cloud Storage, Cloud Pub/Sub, and Cloud Dataflow
D. Cloud Dataproc, Cloud Pub/Sub, Cloud SQL, and Cloud Dataflow
E. Cloud Pub/Sub, Compute Engine, Cloud Storage, and Cloud Dataproc
A is not correct because Cloud SQL is the only storage listed, is limited to 10 TB of storage, and is better suited for transactional workloads. Mountkirk Games needs queries to access at least 30,720 GB of historical data for analytic purposes.
B is correct because:
-Cloud Dataflow dynamically scales up or down, can process data in real time, and is ideal for processing data that arrives late using Beam windows and triggers.
-Cloud Storage can be the landing space for files that are regularly uploaded by users’ mobile devices.
-Cloud Pub/Sub can ingest the streaming data from the mobile users.
BigQuery can query more than 10 TB of historical data.
C is not correct because Cloud SQL is the only storage listed, is limited to 30,720 GB of storage, and is better suited for transactional workloads. Mountkirk Games needs queries to access at least 10 TB of historical data for analytic purposes.
D is not correct because Cloud SQL is limited to 30,720 GB of storage and is better suited for transactional workloads. Mountkirk Games needs queries to access at least 10 TB of historical data for analytics purposes.
E is not correct because Mountkirk Games needs the ability to query historical data. While this might be possible using workarounds, such as BigQuery federated queries for Cloud Storage or Hive queries for Cloud Dataproc, these approaches are more complex. BigQuery is a simpler and more flexible product that fulfills those requirements.
https: //cloud.google.com/sql/docs/quotas#fixed-limits
https: //beam.apache.org/documentation/programming-guide/#windowing
https: //beam.apache.org/documentation/programming-guide/#triggers
https: //cloud.google.com/bigquery/external-data-sources
37
For this question, refer to the Mountkirk Games case study. https://cloud.google.com/certification/guides/cloud-architect/casestudy-mountkirkgames-rev2
Mountkirk Games has deployed their new backend on Google Cloud Platform (GCP). You want to create a thorough testing process for new versions of the backend before they are released to the public. You want the testing environment to scale in an economical way. How should you design the process?
A. Create a scalable environment in Google Cloud for simulating production load.
B. Use the existing infrastructure to test the Google Cloud-based backend at scale.
C. Build stress tests into each component of your application and use resources from the already deployed production backend to simulate load.
D. Create a set of static environments in Google Cloud to test different levels of load—for example, high, medium, and low.
A is correct because simulating production load in Google Cloud can scale in an economical way.
B is not correct because one of the pain points about the existing infrastructure was precisely that the environment did not scale well.
C is not correct because it is a best practice to have a clear separation between test and production environments. Generating test load should not be done from a production environment.
D is not correct because Mountkirk Games wants the testing environment to scale as needed. Defining several static environments for specific levels of load goes against this requirement.
https: //cloud.google.com/community/tutorials/load-testing-iot-using-gcp-and-locust
https: //github.com/GoogleCloudPlatform/distributed-load-testing-using-kubernetes
38
For this question, refer to the Mountkirk Games case study. https://cloud.google.com/certification/guides/cloud-architect/casestudy-mountkirkgames-rev2
Mountkirk Games wants to set up a continuous delivery pipeline. Their architecture includes many small services that they want to be able to update and roll back quickly. Mountkirk Games has the following requirements: (1) Services are deployed redundantly across multiple regions in the US and Europe, (2) Only frontend services are exposed on the public internet, (3) They can reserve a single frontend IP for their fleet of services, and (4) Deployment artifacts are immutable. Which set of products should they use?
A. Cloud Storage, Cloud Dataflow, Compute Engine
B. Cloud Storage, App Engine, Cloud Load Balancing
C. Container Registry, Google Kubernetes Engine, Cloud Load Balancing
D. Cloud Functions, Cloud Pub/Sub, Cloud Deployment Manager
A is not correct because Mountkirk Games wants to set up a continuous delivery pipeline, not a data processing pipeline. Cloud Dataflow is a fully managed service for creating data processing pipelines.
B is not correct because a Cloud Load Balancer distributes traffic to Compute Engine instances. App Engine and Cloud Load Balancer are parts of different solutions.
C is correct because:
-Google Kubernetes Engine is ideal for deploying small services that can be updated and rolled back quickly. It is a best practice to manage services using immutable containers. -Cloud Load Balancing supports globally distributed services across multiple regions. It provides a single global IP address that can be used in DNS records. Using URL Maps, the requests can be routed to only the services that Mountkirk wants to expose. -Container Registry is a single place for a team to manage Docker images for the services.
D is not correct because you cannot reserve a single frontend IP for cloud functions. When deployed, an HTTP-triggered cloud function creates an endpoint with an automatically assigned IP.
https: //cloud.google.com/sql/docs/quotas#fixed-limits
https: //beam.apache.org/documentation/programming-guide/#windowing
https: //beam.apache.org/documentation/programming-guide/#triggers
https: //cloud.google.com/bigquery/external-data-sources
https: //cloud.google.com/solutions/using-apache-hive-on-cloud-dataproc
39
Your customer is moving their corporate applications to Google Cloud. The security team wants detailed visibility of all resources in the organization. You use Resource Manager to set yourself up as the Organization Administrator. Which Cloud Identity and Access Management (Cloud IAM) roles should you give to the security team while following Google recommended practices?
A. Organization viewer, Project owner
B. Organization viewer, Project viewer
C. Organization administrator, Project browser
D. Project owner, Network administrator
A is not correct because Project owner is too broad. The security team does not need to be able to make changes to projects.
B is correct because:
- Organization viewer grants the security team permissions to view the organization's display name.
- Project viewer grants the security team permissions to see the resources within projects.
C is not correct because Organization Administrator is too broad. The security team does not need to be able to make changes to the organization.
D is not correct because Project Owner is too broad. The security team does not need to be able to make changes to projects.
https://cloud.google.com/resource-manager/docs/access-control-org#using_predefined_roles
40
To reduce costs, the Director of Engineering has required all developers to move their development infrastructure resources from on-premises virtual machines (VMs) to Google Cloud. These resources go through multiple start/stop events during the day and require state to persist. You have been asked to design the process of running a development environment in Google Cloud while providing cost visibility to the finance department. Which two steps should you take? (choose two)
A. Use persistent disks to store the state. Start and stop the VM as needed.
B. Use the --auto-delete flag on all persistent disks before stopping the VM.
C. Apply VM CPU utilization label and include it in the BigQuery billing export.
D. Use BigQuery billing export and labels to relate cost to groups.
E. Store all state in a Local SSD, snapshot the persistent disks, and terminate the VM.
A is correct because persistent disks will not be deleted when an instance is stopped.
B is not correct because the --auto-delete flag has no effect unless the instance is deleted. Stopping an instance does not delete the instance or the attached persistent disks.
C is not correct because labels are used to organize instances, not to monitor metrics.
D is correct because exporting daily usage and cost estimates automatically throughout the day to a BigQuery dataset is a good way of providing visibility to the finance department. Labels can then be used to group the costs based on team or cost center.
E is not correct because the state stored in local SSDs will be lost when the instance is stopped.
https: //cloud.google.com/compute/docs/instances/instance-life-cycle
https: //cloud.google.com/sdk/gcloud/reference/compute/instances/set-disk-auto-delete#--auto-delete
https: //cloud.google.com/sdk/gcloud/reference/compute/instances/create#--disk
https: //cloud.google.com/compute/docs/disks/local-ssd#data_persistence
https: //cloud.google.com/billing/docs/how-to/export-data-bigquery
https: //cloud.google.com/resource-manager/docs/creating-managing-labels
41
The database administration team has asked you to help them improve the performance of their new database server running on Compute Engine. The database is used for importing and normalizing the company’s performance statistics. It is built with MySQL running on Debian Linux. They have an n1-standard-8 virtual machine with 80 GB of SSD zonal persistent disk which they can't restart until the next maintenance event. What should they change to get better performance from this system as soon as possible and in a cost-effective manner?
A. Increase the virtual machine’s memory to 64 GB.
B. Create a new virtual machine running PostgreSQL.
C. Dynamically resize the SSD persistent disk to 500 GB.
D. Migrate their performance metrics warehouse to BigQuery.
A is not correct because increasing the memory size requires a VM restart.
B is not correct because the DB administration team is requesting help with their MySQL instance. Migration to a different product should not be the solution when other optimization techniques can still be applied first.
C is correct because persistent disk performance is based on the total persistent disk capacity attached to an instance and the number of vCPUs that the instance has. Incrementing the persistent disk capacity will increment its throughput and IOPS, which in turn improve the performance of MySQL.
D is not correct because the DB administration team is requesting help with their MySQL instance. Migration to a different product should not be the solution when other optimization techniques can still be applied first.
https: //cloud.google.com/compute/docs/disks/performance
https: //cloud.google.com/compute/docs/disks/#pdspecs
42
You have been tasked with interviewing line-of-business owners about their needs for a new
cloud application. Which of the following do you expect to find?
A. A comprehensive list of defined business and technical requirements
B. That their business requirements do not have a one-to-one correlation with technical
requirements
C. Business and technical requirements in conflict
D. Clear consensus on all requirements
B. The correct answer is B. Business requirements are high-level, business-oriented
requirements that are rarely satisfied by meeting a single technical requirement. Option
A is incorrect because business sponsors rarely have sufficient understanding of technical
requirements in order to provide a comprehensive list. Option C is incorrect, because business requirements constrain technical options but should not be in conflict. Option D is
incorrect because there is rarely a clear consensus on all requirements. Part of an architect’s
job is to help stakeholders reach a consensus
43
You have been asked by stakeholders to suggest ways to reduce operational expenses as part
of a cloud migration project. Which of the following would you recommend?
A. Managed services, preemptible machines, access controls
B. Managed services, preemptible machines, autoscaling
C. NoSQL databases, preemptible machines, autoscaling
D. NoSQL databases, preemptible machines, access controls
B. The correct answer is B. Managed services relieve DevOps work, preemptible machines
cost significantly less than standard VMs, and autoscaling reduces the chances of running
unnecessary resources. Options A and D are incorrect because access controls will not help
reduce costs, but they should be used anyway. Options C and D are incorrect because there
is no indication that a NoSQL database should be used
44
Some executives are questioning your recommendation to employ continuous integration/
continuous deployment (CI/CD). What reasons would you give to justify your
recommendation?
A. CI/CD supports small releases, which are easier to debug and enable faster feedback.
B. CI/CD is used only with preemptible machines and therefore saves money.
C. CI/CD fits well with waterfall methodology but not agile methodologies.
D. CI/CD limits the number of times code is released.
A. The correct answer is A. CI/CD supports small releases, which are easier to debug
and enable faster feedback. Option B is incorrect, as CI/CD does not only use preemptible
machines. Option C is incorrect because CI/CD works well with agile methodologies. Option
D is incorrect, as there is no limit to the number of times new versions of code can be released
45
The finance director has asked your advice about complying with a document retention
regulation. What kind of service-level objective (SLO) would you recommend to ensure
that the finance director will be able to retrieve sensitive documents for at least the next
seven years? When a document is needed, the finance director will have up to seven days to
retrieve it. The total storage required will be approximately 100 GB.
A. High availability SLO
B. Durability SLO
C. Reliability SLO
D. Scalability SLO
B. The correct answer is B. The finance director needs to have access to documents for
seven years. This requires durable storage. Option A is incorrect because the access does
not have to be highly available; as long as the finance director can access the document in a
reasonable period of time, the requirement can be met. Option C is incorrect because reliability is a measure of being available to meet workload demands successfully. Option D is
incorrect because the requirement does not specify the need for increasing and decreasing
storage to meet the requirement.
46
You are facilitating a meeting of business and technical managers to solicit requirements for
a cloud migration project. The term incident comes up several times. Some of the business
managers are unfamiliar with this term in the context of IT. How would you describe an
incident?
A. A disruption in the ability of a DevOps team to complete work on time
B. A disruption in the ability of the business managers to approve a project plan on
schedule
Review Questions 21
C. A disruption that causes a service to be degraded or unavailable
D. A personnel problem on the DevOps team
C. The correct answer is C. An incident in the context of IT operations and service reliability is a disruption that degrades or stops a service from functioning. Options A and B are
incorrect—incidents are not related to scheduling. Option D is incorrect; in this context,
incidents are about IT services, not personnel.
47
You have been asked to consult on a cloud migration project that includes moving private
medical information to a storage system in the cloud. The project is for a company in
the United States. What regulation would you suggest that the team review during the
requirements-gathering stages?
A. General Data Protection Regulations (GDPR)
B. Sarbanes–Oxley (SOX)
C. Payment Card Industry Data Security Standard (PCI DSS)
D. Health Insurance Portability and Accountability Act (HIPAA)
D. The correct answer is D. HIPAA governs, among other things, privacy and data protections for private medical information. Option A is incorrect, as GDPR is a European Union
regulation. Option B is incorrect, as SOX is a U.S. financial reporting regulation. Option C
is incorrect, as PCI DSS is a payment card industry regulation.
48
You are in the early stages of gathering business and technical requirements. You have
noticed several references about needing up-to-date and consistent information regarding
product inventory. Inventory is managed on a global scale, and the warehouses storing
inventory are located in North America, Africa, Europe, and India. Which managed
database solution in Google Cloud would you include in your set of options for an
inventory database?
A. Cloud Storage
B. BigQuery
C. Cloud Spanner
D. Microsoft SQL Server
C. The correct answer is C. Cloud Spanner is a globally consistent, horizontally scalable
relational database. Option A is incorrect. Cloud Storage is not a database; rather, it is an
object storage system. Option B is incorrect because BigQuery is an analytics database.
Option D is incorrect, as Microsoft SQL Server is not a managed database in Google Cloud.
49
A developer at Mountkirk Games is interested in how architects decide which database
to use. The developer describes a use case that requires a document store. The developer
would rather not manage database servers or have to run backups. What managed service
would you suggest the developer consider?
A. Cloud Datastore
B. Cloud Spanner
C. Cloud Storage
D. BigQuery
A. The correct answer is A. Cloud Datastore is a managed document database and a good
fit for storing documents. Option B is incorrect because Cloud Spanner is a relational
database and globally scalable. There is no indication that the developer needs a globally scalable solution. Option C is incorrect, as Cloud Storage is an object storage system,
not a managed database. Option D is incorrect because BigQuery is an analytic database
designed for data warehousing and similar applications.
50
Members of your company’s legal team are concerned about using a public cloud service
because other companies, organizations, and individuals will be running their systems in
the same cloud. You assure them that your company’s resources will be isolated and not
network-accessible to others because of what networking resource in Google Cloud?
A. CIDR blocks
B. Direct connections
C. Virtual private clouds
D. Cloud Pub/Sub
C. The correct answer is C. VPCs isolate cloud resources from resources in other VPCs,
unless VPCs are intentionally linked. Option A is incorrect because a CIDR block has
to do with subnet IP addresses. Option B is incorrect, as direct connections are for transmitting data between a data center and Google Cloud—it does not protect resources in
the cloud. Option D is incorrect because Cloud Pub/Sub is a messaging service, not a networking service.
51
. What two business drivers are behind Dress4Win’s interest in moving to the cloud?
A. Insufficient infrastructure capacity and desire to be more agile
B. Insufficient infrastructure and competitors moving to the cloud
C. Competitors moving to the cloud and desire to be more agile
D. Insufficient infrastructure and short-term cost savings
A. The correct answer is A. Dress4Win is at capacity with its existing infrastructure and
wants to innovate faster. Options B and C are incorrect because the decision is not influenced by competitors moving to the cloud. Option D is incorrect because short-term cost
savings are not a consideration.
52
Dress4Win is considering replacing its self-managed MySQL database with a managed
service. Which Google Cloud service would you recommend that they consider?
A. Cloud Dataproc
B. Cloud Dataflow
C. Cloud SQL
D. PostgreSQL
C. The correct answer is C. Cloud SQL offers a managed MySQL service. Options A and
B are incorrect, as neither is a database. Cloud Dataproc is a managed Hadoop and Spark
service. Cloud Dataflow is a stream and batch processing service. Option D is incorrect,
because PostgreSQL is another relational database, but it is not a managed service. PostgreSQL is an option in Cloud SQL, however.
53
Which of the following requirements from a customer makes you think the application
should run in Compute Engine and not App Engine?
A. Dynamically scale up or down based on game activity
B. Connect to a database
C. Run a hardened Linux distro on a virtual machine
D. Don’t lose data
C. The correct answer is C. In Compute Engine, you create virtual machines and choose
which operating system to run. All other requirements can be realized in App Engine.
54
Consider the TerramEarth case study. What aspect of that case study prompts you to
consider potentially significant changes to requirements in the future?
A. Dealers will want more reports about their customers.
B. Of 20 million pieces of equipment, only 200,000 have cellular connections;
19,800,000 additional pieces of equipment may someday transmit data in real time
instead of downloading it in batches.
C. TerramEarth is in a competitive industry.
D. TerramEarth would like to partner with other companies to improve overall service to
their customers
B. The correct answer is B. A significant increase in the use of streaming input will
require changes to how data is ingested and require scalable ingestion services. An
increase of almost two orders of magnitude in the number of pieces of equipment
transmitting data will likely require architectural changes. Option A is incorrect, as
additional reporting is easily accommodated. Option C is incorrect because the initial
design will take into account that TerramEarth is in a competitive industry. Option D is
incorrect, as collaborating with other companies will not require significant changes in
systems design
55
Mountkirk Games wants to store player game data in a time-series database. Which Google
Cloud managed database would you recommend?
A. Bigtable
B. BigQuery
C. Cloud Storage
D. Cloud Dataproc
A. The correct answer is A. Cloud Bigtable is a scalable, wide-column database designed
for low-latency writes, making it a good choice for time-series data. Option B is incorrect because BigQuery is an analytic database not designed for the high volume of
low-latency writes that will need to be supported. Options C and D are not managed
databases.
56
The game analytics platform for Mountkirk Games requires analysts to be able to query up
to 10 TB of data. What is the best managed database solution for this requirement?
A. Cloud Spanner
B. BigQuery
C. Cloud Storage
D. Cloud Dataprep
B. The correct answer is B. This is a typical use case for BigQuery, and it fits well with its
capabilities as an analytic database. Option A is incorrect, as Cloud Spanner is best used
for transaction processing on a global scale. Options C and D are not managed databases.
Cloud Storage is an object storage service; Cloud Dataprep is a tool for preparing data
for analysis.
57
In the Dress4Win case study, the volume of data and compute load will grow with respect
to what factor?
A. The number of customers, designers, and retailers
B. The time the application is running
C. The type of storage used
D. Compliance with regulations
A. Option A is correct. Dress4Win is a consumer, e-commerce service that will grow
with respect to the number of customers. Also, the number of designers and retailers will
influence the growth in demand for compute and storage resources. Option B is incorrect
because the length of run time is not relevant to compute or storage requirements. The type
of storage used does not influence the amount of data the application needs to manage, or
the amount of computing resources needed. Compliance and regulations may have some
effect on security controls and monitoring, but it will not influence compute and storage
resources in a significant way
58
You have received complaints from customers about long wait times while loading application pages in their browsers, especially pages with several images. Your director has tasked
you with reducing latency when accessing and transmitting data to a client device outside
the cloud. Which of the following would you use? (Choose two.)
A. Multiregional storage
B. Coldline storage
C. CDN
D. Cloud Pub/Sub
E. Cloud Dataflow
A, C. Options A and C are correct. Both multiregional cloud storage and CDNs distribute
data across a geographic area. Option B is incorrect because Coldline storage is used for
archiving. Option D is incorrect because Cloud Pub/Sub is a messaging queue, not a storage
system. Option E is a managed service for batch and stream processing.
59
Mountkirk Games will analyze game players’ usage patterns. This will require collecting
time-series data including game state. What database would be a good option for doing
this?
A. BigQuery
B. Bigtable
C. Cloud Spanner
D. Cloud Storage
B. Option B is correct. High volumes of time-series data need low-latency writes and scalable storage. Time-series data is not updated after it is collected. This makes Bigtable, a
wide-column data store with low-latency writes, the best option. Option A is wrong because
BigQuery is an analytic database designed for data warehousing. Option C is wrong
because Cloud Spanner is a global relational database. Write times would not be as fast as
they would be using Bigtable, and the use case does not take advantage of Cloud Spanner’s
strong consistency in a horizontally scalable relational database. Option D is not a good
option because it is an object store, and it is not designed for large volumes of individual
time-series data points.
60
You have been hired to consult with a new data warehouse team. They are struggling to
meet schedules because they repeatedly find problems with data quality and have to write
preprocessing scripts to clean the data. What managed service would you recommend for
addressing these problems?
A. Cloud Dataflow
B. Cloud Dataproc
C. Cloud Dataprep
D. Cloud Datastore
A. Option A is correct. Cloud Dataflow is a batch and stream processing service that can be
used for transforming data before it is loaded into a data warehouse. Option C is incorrect,
Cloud Dataprep is used to prepare data for analysis and machine learning. Option B, Cloud
Dataproc, is a managed Hadoop and Spark service, not a data cleaning and preparing service. Option D, Cloud Datastore, is a document database, not a data processing service.
61
You have deployed an application that receives data from sensors on TerramEarth
equipment. Sometimes more data arrives than can be processed by the current set of
Compute Engine instances. Business managers do not want to run additional VMs. What
changes could you make to ensure that data is not lost because it cannot be processed as it
is sent from the equipment? Assume that business managers want the lowest-cost solution.
A. Write data to local SSDs on the Compute Engine VMs.
B. Write data to Cloud Memorystore, and have the application read data from the cache.
C. Write data from the equipment to a Cloud Pub/Sub queue, and have the application
read data from the queue.
D. Tune the application to run faster.
C. The correct answer is C, write data to a Cloud Pub/Sub topic. The data can accumulate
there as the application processes the data. No data is lost because Pub/Sub will scale as
needed. Option A is not a good option because local storage does not scale. Option B is not a
good choice because caches are used to provide low-latency access to data that is frequently
accessed. Cloud Memorystore does not scale as well as Cloud Pub/Sub, and it may run out of
space. Option D is not a good choice because tuning will require developers to invest potentially significant amounts of time without any guarantee of solving the problem. Also, even
with optimizations, even larger spikes in data ingestion could result in the same problem of
the processing application not being able to keep up with the rate at which data is arriving
62
Your company uses Apache Spark for data science applications. Your manager has asked
you to investigate running Spark in the cloud. Your manager’s goal is to lower the overall
cost of running and managing Spark. What would you recommend?
A. Run Apache Spark in Compute Engine.
B. Use Cloud Dataproc with preemptible virtual machines.
C. Use Cloud Dataflow with preemptible virtual machines.
D. Use Cloud Memorystore with Apache Spark running in Compute Engine.
B. Option B is correct. Using Cloud Dataproc will reduce the costs of managing the Spark
cluster, while using preemptible VMs will reduce the compute charges. Option A is not
the best option because you will have to manage the Spark cluster yourself, which will
increase the total cost of ownership. Option C is incorrect as Cloud Dataflow is not a
managed Spark service. Option D is incorrect because Cloud Memorystore does not reduce
the cost of running Apache Spark and managing a cluster in Compute Engine is not the
most cost-effective
63
You are working with a U.S. hospital to extract data from an electronic health record
(EHR) system. The hospital has offered to provide business requirements, but there is little
information about regulations in the documented business requirements. What regulations
would you look to for more guidance on complying with relevant regulations?
A. GDPR
B. SOX
C. HIPAA
D. PCI DSS
C. The relevant health regulation is HIPAA, which regulates healthcare data in the United
States. Option A is incorrect, as GDPR is a European Union privacy regulation. Option
B is incorrect, as SOX is a regulation that applies to the financial industry. Option D is
incorrect, because the Payment Card Industry Data Security Standard does not apply to
healthcare data.
64
```
What security control can be used to help detect changes to data?
A. Firewall rules
B. Message digests
C. Authentication
D. Authorization
```
B. Option B is correct. Message digests are used to detect changes in files. Option A is
incorrect because firewall rules block network traffic and are not related to detecting
changes to data. Options C and D are important for controlling access to data, but they are
not directly related to detecting changes to data.
65
Your company has a data classification scheme for categorizing data as secret, sensitive,
private, and public. There are no confidentiality requirements for public data. All other
data must be encrypted at rest. Secret data must be encrypted with keys that the company
controls. Sensitive and private data can be encrypted with keys managed by a third
party. Data will be stored in GCP. What would you recommend in order to meet these
requirements while minimizing cost and administrative overhead?
A. Use Cloud KMS to manage keys for all data.
B. Use Cloud KMS for secret data and Google default encryption for other data.
C. Use Google default encryption for all data.
D. Use a custom encryption algorithm for all data.
B. B is correct. Cloud KMS allows the customer to manage keys used to encrypt secret
data. The requirements for the other categories are met by GCP’s default encryption-at-rest
practice. Public data does not need to be encrypted, but there is no additional cost or overhead for having it encrypted at rest. Option A would meet the security requirements, but
it would involve managing keys for more data than is necessary, and that would increase
administrative overhead. Option C does not meet the requirements of secret data. Option D
is a terrible choice. Encryption algorithms are difficult to develop and potentially vulnerable
to cryptanalysis attacks. It would cost far more to develop a strong encryption algorithm
than to use Cloud KMS and default encryption.
66
You manage a service with several databases. The queries to the relational database are
increasing in latency. Reducing the amount of data in tables will improve performance
and reduce latency. The application administrator has determined that approximately
60 percent of the data in the database is more than 90 days old and has never been queried
and does not need to be in the database. You are required to keep the data for five years in
case it is requested by auditors. What would you propose to decrease query latency without
increasing costs—or at least keeping any cost increases to a minimum?
A. Horizontally scale the relational database.
B. Vertically scale the relational database.
C. Export data more than 90 days old, store it in Cloud Storage Coldline class storage,
and delete that data from the relational database.
D. Export data more than 90 days old, store it in Cloud Storage multiregional class
storage, and delete that data from the relational database.
C. The correct answer is C. Data that is not queried does not need to be in the database to
meet business requirements. If the data is needed, it can be retrieved from other storage systems, such as Cloud Storage. Exporting and deleting data will reduce the amount of data in
tables and improve performance. Since the data is rarely accessed, it is a good candidate for
archival, Coldline storage. Answers A and B are incorrect because scaling either vertically
or horizontally will increase costs more than the cost of storing the data in archival storage.
Option D is incorrect because multiregional storage is more expensive than Coldline storage and multiregion access is not needed.
67
Your company is running several custom applications that were written by developers who
are no longer with the company. The applications frequently fail. The DevOps team is paged
more for these applications than any others. You propose replacing those applications with
several managed services in GCP. A manager who is reviewing your cost estimates for using
managed services in GCP notes that the cost of the managed services will be more than what
they pay for internal servers. What would you recommend as the next step for the manager?
A. Nothing. The manager is correct—the costs are higher. You should reconsider your
recommendation.
B. Suggest that the manager calculate total cost of ownership, which includes the cost to
support the applications as well as infrastructure costs.
C. Recommend running the custom applications in Compute Engine to lower costs.
D. Recommend rewriting the applications to improve reliability.
B. Option B is correct. The manager does not have an accurate cost estimate of supporting
the applications if operational support costs are not considered. The manager should have
an accurate estimate of TCO before proceeding. Option A is incorrect because the manager
does not have an accurate estimate of all costs. Option C is incorrect because it does not
address the reliability issues with the applications. Option D may be a reasonable option,
but if managed services meet the requirements, using them will solve the reliability issues
faster than developing new applications.
68
A director at Mountkirk Games has asked for your recommendation on how to measure
the success of the migration to GCP. The director is particularly interested in customer
satisfaction. What KPIs would you recommend?
A. Average revenue per customer per month
B. Average time played per customer per week
C. Average time played per customer per year
D. Average revenue per customer per year
B. Option B is the best answer because it is a measure of how much customers are engaged
in the game and playing. If average time played goes down, this is an indicator that customers are losing interest in the game. If the average time played goes up, they are more
engaged and interested in the game. Options A and D are incorrect because revenue does
not necessarily correlate with customer satisfaction. Also, it may not correlate with how
much customers played the game if revenue is based on monthly subscriptions, for example.
Option C is wrong because a year is too long a time frame for detecting changes as rapidly
as one can with a weekly measure.
69
Mountkirk Games is implementing a player analytics system. You have been asked to
document requirements for a stream processing system that will ingest and preprocess data
before writing it to the database. The preprocessing system will collect data about each
player for one minute and then write a summary of statistics about that database. The project
manager has provided the list of statistics to calculate and a rule for calculating values for
missing data. What other business requirements would you ask of the project manager?
A. How long to store the data in the database?
B. What roles and permissions should be in place to control read access to data in the
database?
C. How long to wait for late-arriving data?
D. A list of managed services that can be used in this project
C. Option C is correct. In stream processing applications that collect data for a time and
then produce summary or aggregated data, there needs to be a limit on how long the processor waits for late-arriving data before producing results. Options A and B are incorrect
because you do not need to know requirements for data lifecycle management or access
controls to the database at this point, since your focus is on ingesting raw data and writing
statistics to the database. Option D is incorrect. An architect should provide that list to a
project manager, not the other way around.
70
A new data warehouse project is about to start. The data warehouse will collect data
from 14 different sources initially, but this will likely grow over the next 6 to 12 months.
What managed GCP service would you recommend for managing metadata about the data
warehouse sources?
A. Data Catalog
B. Cloud Dataprep
C. Cloud Dataproc
D. BigQuery
A. The correct option is A. Data Catalog is a managed service for metadata. Option B
is incorrect, as Dataprep is a tool for preparing data for analysis and machine learning.
Option C is incorrect, as Dataproc is a managed Hadoop and Spark service. Option D is
incorrect because BigQuery is a database service designed for analytic databases and data
warehousing
71
You are consulting for a multinational company that is moving its inventory system to GCP. The
company wants to use a managed database service, and it requires SQL and strong consistency.
The database should be able to scale to global levels. What service would you recommend?
A. Bigtable
B. Cloud Spanner
C. Cloud Datastore
D. BigQuery
B. The correct option is B. Cloud Spanner is a horizontally scalable relational database that
provides strong consistency, SQL, and scales to a global level. Options A and C are incorrect because they do not support SQL. Option D is incorrect because an inventory system
is a transaction processing system, and BigQuery is designed for analytic, not transaction
processing systems.
72
TerramEarth has interviewed dealers to better understand their needs regarding data. Dealers would like to have access to the latest data available, and they would like to minimize
the amount of data they have to store in their databases and object storage systems. How
would you recommend that TerramEarth provide data to their dealers?
A. Extract dealer data to a CSV file once per night during off-business hours and upload
it to a Cloud Storage bucket accessible to the dealer.
B. Create an API that dealers can use to retrieve specific pieces of data on an as-needed
basis.
C. Create a database dump using the database export tool so that dealers can use the
database import tool to load the data into their databases.
D. Create a user account on the database for each dealer and have them log into the
database to run their own queries.
B. Option B is correct. An API would allow dealers to access up-to-date information
and allow them to query only for the data that they need. Dealers do not need to know
implementation details of TerramEarth’s database. Options A and C are incorrect because
nightly extracts or exports would not give access to up-to-date data, which could change
during the day. Option D is incorrect because it requires the dealers to understand how to
query a relational database. Also, it is not a good practice to grant direct access to important business databases to people or services outside the company.
73
Your company has large volumes of unstructured data stored on several network-attached
storage systems. The maintenance costs are increasing, and management would like to
consider alternatives. What GCP storage system would you recommend?
A. Cloud SQL
B. Cloud Storage
C. Cloud Datastore
D. Bigtable
B. The correct option is B. Cloud Storage is an object storage system well suited to storing unstructured data. Option A is incorrect because Cloud SQL provides relational databases that are used for structured data. Option C is incorrect because Cloud Datastore is a
NoSQL document database used with flexible schema data. Option D is incorrect, as Bigtable is a wide-column database that is not suitable for unstructured data.
74
A customer-facing application is built using a microservices architecture. One of the
services does not scale as fast as the service that sends it data. This causes the sending
service to wait while the other service processes the data. You would like to change the
integration to use asynchronous instead of synchronous calls. What is one way to do this?
A. Create a Cloud Pub/Sub topic, have the sending service write data to the topic, and
have the receiving service read from the topic.
B. Create a Cloud Storage bucket, have the sending service write data to the topic, and
have the receiving service read from the topic.
C. Have the sending service write data to local drives, and have the receiving service read
from those drives.
D. Create a Bigtable database, have the sending service write data to the topic, and have
the receiving service read from the topic.
A. Option A is correct. Cloud Pub/Sub is designed to provide messaging services and fits
this use case well. Options B and D are incorrect because, although you may be able to
implement asynchronous message exchange using those storage systems, it would be inefficient and require more code than using Cloud Pub/Sub. Option C is incorrect because this
would require both the sending and receiving services to run on the same VM.
75
A product manager at TerramEarth would like to use the data that TerramEarth collects to
predict when equipment will break down. What managed services would you recommend
TerramEarth to consider?
A. Bigtable
B. Cloud Dataflow
C. Cloud AutoML
D. Cloud Spanner
C. The correct answer is C. Cloud AutoML is a managed service for building machine
learning models. TerramEarth’s data could be used to build a predictive model using
AutoML. Options A and D are incorrect—they are databases and do not have the tools
for building predictive models. Option B is wrong because Cloud Dataflow is a stream and
batch processing service.
76
You are advising a customer on how to improve the availability of a data storage solution.
Which of the following general strategies would you recommend?
A. Keeping redundant copies of the data
B. Lowering the network latency for disk writes
C. Using a NoSQL database
D. Using Cloud Spanner
A. The correct answer is A. Redundancy is a general strategy for improving availability.
Option B is incorrect because lowering network latency will not improve availability of
the data storage system. Options C and D are incorrect because there is no indication that
either a NoSQL or a relational database will meet the overall storage requirements of the
system being discussed.
77
A team of data scientists is analyzing archived data sets. The model building procedures
run in batches. If the model building system is down for up to 30 minutes per day, it does
not adversely impact the data scientists’ work. What is the minimal percentage availability
among the following options that would meet this requirement?
A. 99.99 percent
B. 99.90 percent
C. 99.00 percent
D. 99.999 percent
C. The minimum percentage availability that meets the requirements is option C, which
allows for up to 14.4 minutes of downtime per day. All other options would allow for less
downtime, but that is not called for by the requirements.
78
Your development team has recently triggered three incidents that resulted in service
disruptions. In one case, an engineer mistyped a number in a configuration file and in the
other cases specified an incorrect disk configuration. What practices would you recommend
to reduce the risk of these types of errors?
A. Continuous integration/continuous deployment
B. Code reviews of configuration files
C. Vulnerability scanning
D. Improved access controls
B. The correct answer is B. A code review is a software engineering practice that requires
an engineer to review code with another engineer before deploying it. Option A would
not solve the problem, as continuous integration reduces the amount of effort required to
deploy new versions of software. Options C and D are both security controls, which would
not help identify misconfigurations.
79
Your company is running multiple VM instances that have not had any downtime in the
past several weeks. Recently, several of the physical servers suffered disk failures. The
applications running on the servers did not have any apparent service disruptions. What
feature of Compute Engine enabled that?
A. Preemptible VMs
B. Live migration
C. Canary deployments
D. Redundant array of inexpensive disks
B. The correct answer is B, Live Migration, which moves running VMs to different
physical servers without interrupting the state of the VM. Option A is incorrect because
preemptible VMs are low-cost VMs that may be taken back by Google at any time.
Option C is incorrect, as canary deployments are a type of deployment—not a feature
of Compute Engine. Option D is incorrect, as arrays of disks are not directly involved in
preserving the state of a VM and moving the VM to a functioning physical server.
80
You have deployed an application on an instance group. The application is not functioning
correctly. What is a possible outcome?
A. The application shuts down when the instance group time-to-live (TTL) threshold is
reached.
B. The application shuts down when the health check fails.
C. The VM shuts down when the instance group TTL threshold is reached and a new VM
is started.
D. The VM shuts down when the health check fails and a new VM is started
D. Option D is correct. When a health check fails, the failing VM is replaced by a new VM
that is created using the instance group template to configure the new VM. Options A and
C are incorrect, as TTL is not used to detect problems with application functioning. Option
B is incorrect because the application is not shut down when a health check fails.
81
Mountkirk Games is growing its user base in North America, Europe, and Asia. Executives
are concerned that players in Europe and Asia will have a degraded experience if the game
backend runs only in North America. What would you suggest as a way to improve latency
and game experience for users in Europe and Asia?
A. Use Cloud Spanner to have a globally consistent, horizontally scalable relational database.
B. Create instance groups running the game backend in multiple regions across North
America, Europe, and Asia. Use global load balancing to distribute the workload.
C. Use Standard Tier networking to ensure that data sent between regions is routed over
the public Internet.
D. Use a Cloud Memorystore cache in front of the database to reduce database read latency.
B. The correct answer is B. Creating instance groups in multiple regions and routing
workload to the closest region using global load balancing will provide the most consistent experience for users in different geographic regions. Option A is incorrect because
Cloud Spanner is a relational database and does not affect how game backend services
are run except for database operations. Option C is incorrect, as routing traffic over the
public Internet means traffic will experience the variance of public Internet routes between
regions. Option D is incorrect. A cache will reduce the time needed to read data, but it
will not affect network latency when that data is transmitted from a game backend to the
player’s device.
82
What configuration changes are required to ensure high availability when using Cloud
Storage or Cloud Filestore?
A. A sufficiently long TTL must be set.
B. A health check must be specified.
C. Both a TTL and health check must be specified.
D. Nothing. Both are managed services. GCP manages high availability.
D. The correct answer is D. Users do not need to make any configuration changes when
using Cloud Storage or Cloud Filestore. Both are fully managed services. Options A and C
are incorrect because TTLs do not need to be set to ensure high availability. Options B and
C are incorrect because users do not need to specify a health check for managed storage
services.
83
The finance director in your company is frustrated with the poor availability of an
on-premises finance data warehouse. The data warehouse uses a commercial relational
database that only scales by buying larger and larger servers. The director asks for your
advice about moving the data warehouse to the cloud and if the company can continue to
use SQL to query the data warehouse. What GCP service would you recommend to replace
the on-premises data warehouse?
A. Bigtable
B. BigQuery
C. Cloud Datastore
D. Cloud Storage
B. The best answer is B. BigQuery is a serverless, fully managed analytic database that
uses SQL for querying. Options A and C are incorrect because both Bigtable and Cloud
Datastore are NoSQL databases. Option D, Cloud Storage, is not a database, and it does
not meet most of the requirements listed.
84
TerramEarth has determined that it wants to use Cloud Bigtable to store equipment
telemetry data transmitted over their cellular network. They have also concluded that they
want two clusters in different regions. Both clusters should be able to respond to read and
write requests. What kind of replication should be used?
A. Primary–hot primary
B. Primary–warm primary
C. Primary–primary
D. Primary read–primary write
C. The correct answer is C. Primary-primary replication keeps both clusters synchronized
with write operations so that both clusters can respond to queries. Options A, B, and D are
not actual replication options.
85
Your company is implementing a hybrid cloud computing model. Line-of-business
owners are concerned that data stored in the cloud may not be available to on-premises
applications. The current network connection is using a maximum of 40 percent of
bandwidth. What would you suggest to mitigate the risk of that kind of service failure?
A. Configure firewall rules to improve availability.
B. Use redundant network connections between the on-premises data center and Google
Cloud.
C. Increase the number of VMs allowed in Compute Engine instance groups.
D. Increase the bandwidth of the network connection between the data center and Google
Cloud.
B. Option B is correct. A redundant network connection would mitigate the risk of losing
connectivity if a single network connection went down. Option A is incorrect, as firewall
rules are a security control and would not mitigate the risk of network connectivity failures.
Option C may help with compute availability, but it does not improve network availability.
Option D does not improve availability, and additional bandwidth is not needed.
86
. A team of architects in your company is defining standards to improve availability. In
addition to recommending redundancy and code reviews for configuration changes, what
would you recommend to include in the standards?
A. Use of access controls
B. Use of managed services for all compute requirements
C. Use of Stackdriver monitoring to alert on changes in application performance
D. Use of Bigtable to collect performance monitoring data
C. The correct answer is C. Stackdriver should be used to monitor applications and infrastructure to detect early warning signs of potential problems with applications or infrastructure. Option A is incorrect because access controls are a security control and not
related to directly improving availability. Option B is incorrect because managed services
may not meet all requirements and so should not be required in a company’s standards.
Option D is incorrect because collecting and storing performance monitoring data does not
improve availability.
87
Why would you want to run long-running, compute-intensive backend computation in a
different managed instance group than on web servers supporting a minimal user interface?
A. Managed instance groups can run only a single application.
B. Managed instance groups are optimized for either compute or HTTP connectivity.
C. Compute-intensive applications have different scaling characteristics from those of
lightweight user interface applications.
D. There is no reason to run the applications in different managed instance groups
C. The correct answer is C. The two applications have different scaling requirements.
The compute-intensive backend may benefit from VMs with a large number of CPUs that
would not be needed for web serving. Also, the frontend may be able to reduce the number of instances when users are not actively using the user interface, but long compute jobs
may still be running in the background. Options A and B are false statements. Option D is
incorrect for the reasons explained in reference to Option C.
88
An instance group is adding more VMs than necessary and then shutting them down. This
pattern is happening repeatedly. What would you do to try to stabilize the addition and
removal of VMs?
A. Increase the maximum number of VMs in the instance group.
B. Decrease the minimum number of VMs in the instance group.
C. Increase the time autoscalers consider when making decisions.
D. Decrease the time autoscalers consider when making decisions
C. The correct answer is C. The autoscaler may be adding VMs because it has not waited
long enough for recently added VMs to start and begin to take on load. Options A and B
are incorrect because changing the minimum and maximum number of VMs in the group
does not affect the rate at which VMs are added or removed. Option D is incorrect because
it reduces the time available for new instances to start taking on workload, so it may actually make the problem worse.
89
Dress4Win has just developed a new feature for its social networking service. Customers
can upload images of their clothes, create montages from those images, and share them
on social networking sites. Images are temporarily saved to locally attached drives as
the customer works on the montage. When the montage is complete, the final version is
copied to a Cloud Storage bucket. The services implementing this feature run in a managed
instance group. Several users have noted that their final montages are not available even
though they saved them in the application. No other problems have been reported with the
service. What might be causing this problem?
A. The Cloud Storage bucket is out of storage.
B. The locally attached drive does not have a filesystem.
C. The users experiencing the problem were using a VM that was shut down by an
autoscaler, and a cleanup script did not run to copy the latest version of the montage to
Cloud Storage.
D. The network connectivity between the VMs and Cloud Storage has failed
C. The correct answer is C. If the server is shut down without a cleanup script, then data
that would otherwise be copied to Cloud Storage could be lost when the VM shuts down.
Option A is incorrect because buckets do not have a fixed amount of storage. Option B is
incorrect because, if it were true, the service would not function for all users—not just several of them. Option D is incorrect because if there was a connectivity failure between the
VM and Cloud Storage, there would be more symptoms of such a failure.
90
Kubernetes uses several abstractions to model and manage computation and applications.
What is the progression of abstractions from the lowest to the highest level ?
A. Pods → Deployments → Services
B. Pods → Services → Deployments
C. Deployments → Services → Pods
D. Deployments → Pods → Services
A. The correct answer is A. Pods are the lowest level of the computation abstractions.
Deployments are collections of pods running a version of an application. Services are sets
of deployments running an application, possibly with multiple versions running in different
deployments. Options B, C, and D are all incorrect in the order of progression from lowest
to highest level of abstraction.
91
Your development team has implemented a new application using a microservices
architecture. You would like to minimize DevOps overhead by deploying the services in a
way that will autoscale. You would also like to run each microservice in containers. What is
a good option for implementing these requirements in Google Cloud Platform?
A. Run the containers in Cloud Functions.
B. Run the containers in Kubernetes Engine.
C. Run the containers in Cloud Dataproc.
D. Run the containers in Cloud Dataflow.
B. The correct answer is B. The requirements are satisfied by the Kubernetes container
orchestration capabilities. Option A is incorrect, as Cloud Functions do not run containers. Option C is incorrect because Cloud Dataproc is a managed service for Hadoop and
Spark. Option D is incorrect, as Cloud Dataflow is a managed service for stream and batch
processing using the Apache Beam model
92
TerramEarth is considering building an analytics database and making it available to
equipment designers. The designers require the ability to query the data with SQL. The
analytics database manager wants to minimize the cost of the service. What would you
recommend?
A. Use BigQuery as the analytics database, and partition the data to minimize the amount
of data scanned to answer queries.
B. Use Bigtable as the analytics database, and partition the data to minimize the amount
of data scanned to answer queries.
C. Use BigQuery as the analytics database, and use data federation to minimize the
amount of data scanned to answer queries.
D. Use Bigtable as the analytics database, and use data federation to minimize the amount
of data scanned to answer queries.
A. The correct answer is A. BigQuery should be used for an analytics database. Partitioning
allows the query processor to limit scans to partitions that might have the data selected in
a query. Options B and D are incorrect because Bigtable does not support SQL. Options C
and D are incorrect because federation is a way of making data from other sources available
within a database—it does not limit the data scanned in the way that partitioning does.
93
Line-of-business owners have decided to move several applications to the cloud. They
believe the cloud will be more reliable, but they want to collect data to test their hypothesis.
What is a common measure of reliability that they can use?
A. Mean time to recovery
B. Mean time between failures
C. Mean time between deployments
D. Mean time between errors
B. The correct answer is B. Mean time between failures is a measure of reliability. Option A
is a measure of how long it takes to recover from a disruption. Options C and D are incorrect because the time between deployments or errors is not directly related to reliability.
94
A group of business executives and software engineers are discussing the level of risk that
is acceptable for a new application. Business executives want to minimize the risk that the
service is not available. Software engineers note that the more developer time dedicated to
reducing risk of disruption, the less time they have to implement new features. How can
you formalize the group’s tolerance for risk of disruption?
A. Request success rate
B. Uptime of service
C. Latency
D. Throughput
A. The correct answer is A. Request success rate is a measure of how many requests were
successfully satisfied. Option B is incorrect because at least some instances of an application
may be up at any time, so it does not reflect the capacity available. Options C and D are not
relevant measures of risk.
95
Your DevOps team recently determined that it needed to increase the size of persistent disks
used by VMs running a business-critical application. When scaling up the size of available
persistent storage for a VM, what other step may be required?
A. Adjusting the filesystem size in the operating system
B. Backing up the persistent disk before changing its size
C. Changing the access controls on files on the disk
D. Update disk metadata, including labels
A. The correct answer is A. The persistent storage may be increased in size, but the operating system may need to be configured to use that additional storage. Option B is incorrect
because while backing up a disk before operating on it is a good practice, it is not required.
Option C is incorrect because changing storage size does not change access control rules.
Option D is incorrect because any disk metadata that needs to change when the size
changes is updated by the resize process.
