Try each one before opening the answer. Every option is explained, with the Google Cloud documentation page that proves it.
Question 1 · choose 1
Log archives land in a Standard-class bucket. They are read often in the first 90 days, almost never after that, and must be deleted after one year. How should you configure the bucket to keep costs down without scripts?
AEnable Object Versioning so older log files move to a cheaper storage class automatically
BCreate the bucket in Coldline from the start and add a Delete rule at age 365 days
CSet a retention policy of 365 days on the bucket so objects expire after one year
DAdd lifecycle rules: SetStorageClass to Coldline at age 90 days, and Delete at age 365 days
Show the answer and why
AEnable Object Versioning so older log files move to a cheaper storage class automatically
Incorrect
Object Versioning has no setting that moves objects to another storage class.
BCreate the bucket in Coldline from the start and add a Delete rule at age 365 days
Incorrect
Coldline's retrieval charges would apply to the frequent reads of the first 90 days.
CSet a retention policy of 365 days on the bucket so objects expire after one year
Incorrect
A retention policy neither deletes objects nor changes their class, so it meets neither cost goal.
DAdd lifecycle rules: SetStorageClass to Coldline at age 90 days, and Delete at age 365 days
Correct
Lifecycle rules apply actions by condition, such as age, so objects move to a cheaper class and are deleted on schedule.
Object Lifecycle Management turns an access pattern into rules: change the class when data cools and delete it when it expires.
A Cloud SQL for MySQL instance holds order data. If someone deletes rows by mistake, the team must be able to restore the database to the minute before the delete at any time in the last seven days. Which TWO settings must be enabled? (Choose TWO.)
AAutomated daily backups of the instance
BHigh availability with a standby in another zone
CA weekly export of the database to Cloud Storage
DA cross-region read replica of the instance
EPoint-in-time recovery, which uses binary logging
Show the answer and why
AAutomated daily backups of the instance
Correct
Point-in-time recovery starts from a backup, so automated backups must be on.
BHigh availability with a standby in another zone
Incorrect
The standby receives the same changes, including the delete, so it protects against zone failure rather than bad writes.
CA weekly export of the database to Cloud Storage
Incorrect
An export captures one moment a week, not any minute within the last seven days.
DA cross-region read replica of the instance
Incorrect
A replica applies the same delete shortly after the primary does, so the rows are gone there too.
EPoint-in-time recovery, which uses binary logging
Correct
Binary logs record every change after a backup, which lets you recover to a specific moment.
Recovering to an exact moment needs a base backup plus a log of changes after it. Replicas and standbys copy mistakes as faithfully as good writes.
Customers who have no Google accounts must download their monthly invoice PDFs from a private Cloud Storage bucket. Each link must work for only 24 hours and only for that customer's file. What should your application generate?
AAn access control list on each object that names the customer's email address
BA signed URL for the object with a 24-hour expiration
CA grant of Storage Object Viewer to allUsers on the bucket, removed after 24 hours
DAn HMAC key for each customer with access to the bucket
Show the answer and why
AAn access control list on each object that names the customer's email address
Incorrect
ACLs grant access to Google identities, which these customers do not have.
BA signed URL for the object with a 24-hour expiration
Correct
A signed URL grants time-limited access to one object to anyone who has the URL, without a Google account.
CA grant of Storage Object Viewer to allUsers on the bucket, removed after 24 hours
Incorrect
This makes every invoice in the bucket public to anyone during the window.
DAn HMAC key for each customer with access to the bucket
Incorrect
HMAC keys are long-lived credentials for applications using the XML API, not short-lived links for people.
Signed URLs embed the authorization in the link: one object, a chosen method and an expiry. They suit sharing with people outside your identity system.
Compliance requires that new objects in a bucket in europe-west3 be encrypted with a Cloud KMS key the company controls. You created a key ring and key in us-central1 and set it as the bucket's default key, but uploads fail. What should you do?
ATurn on Object Versioning so that objects are rewritten with the new key after each upload
BGrant the uploading users the Cloud KMS CryptoKey Encrypter/Decrypter role on the us-central1 key
CSupply a customer-supplied encryption key with each upload instead of a Cloud KMS key
DUse a key in europe-west3 and grant the Cloud Storage service agent Encrypter/Decrypter on it
Show the answer and why
ATurn on Object Versioning so that objects are rewritten with the new key after each upload
Incorrect
Versioning keeps older versions of objects; it has nothing to do with which key encrypts them.
BGrant the uploading users the Cloud KMS CryptoKey Encrypter/Decrypter role on the us-central1 key
Incorrect
Cloud Storage uses the key through its service agent, not with the users' permissions, and the key's location still does not match.
CSupply a customer-supplied encryption key with each upload instead of a Cloud KMS key
Incorrect
Customer-supplied keys work, but then the company must manage and send key material on every request, which is not what was asked.
DUse a key in europe-west3 and grant the Cloud Storage service agent Encrypter/Decrypter on it
Correct
The key must be in a location compatible with the bucket, and the Cloud Storage service agent needs permission to use the key.
With CMEK, a Google-managed service agent encrypts and decrypts on your behalf. It needs the Encrypter/Decrypter role on the key, and the key ring must be in a location that matches the bucket.
A bucket holds internal reports. You want to make sure nobody can ever grant public access to it, even by mistake, while normal IAM access for employees keeps working. What should you enable?
ASigned URLs for every employee
BA firewall rule that blocks internet traffic to the bucket
CObject Versioning on the bucket
DPublic access prevention on the bucket
Show the answer and why
ASigned URLs for every employee
Incorrect
These do not stop anyone from granting public access to the bucket.
BA firewall rule that blocks internet traffic to the bucket
Incorrect
VPC firewall rules do not apply to Cloud Storage access.
CObject Versioning on the bucket
Incorrect
Versioning keeps older versions; it does not stop public grants.
DPublic access prevention on the bucket
Correct
With public access prevention enforced, the bucket and its objects cannot be made public, while other access works as before.
Public access prevention is a guardrail on top of IAM: it blocks allUsers and allAuthenticatedUsers grants on the bucket.
Before approving a project, finance wants an estimate of the monthly cost of 120 TB in Cloud Storage and 15 TB in Filestore, with the estimate shared as a link. Nothing has been created yet. What should you use?
AThe Google Cloud Pricing Calculator
BA Cloud Monitoring dashboard of storage usage
CThe Cloud Asset Inventory console
DThe Budgets page of the billing account
Show the answer and why
AThe Google Cloud Pricing Calculator
Correct
The calculator builds cost estimates for planned products and lets you share them.
BA Cloud Monitoring dashboard of storage usage
Incorrect
Dashboards show usage of existing resources, not future cost.
CThe Cloud Asset Inventory console
Incorrect
Asset Inventory lists existing resources, not prices of planned ones.
DThe Budgets page of the billing account
Incorrect
Budgets track actual spend after resources exist.
Estimate before you build with the Pricing Calculator; track after you build with billing tools.
A bucket's objects are protected with a customer-managed key in Cloud KMS. An administrator disables the key version that encrypted them. What happens to those objects?
AThey are re-encrypted with a Google-managed key automatically
BThey cannot be accessed until the key version is enabled again
CThey stay readable because Google keeps a plaintext copy
DThey are deleted immediately and permanently
Show the answer and why
AThey are re-encrypted with a Google-managed key automatically
Incorrect
Cloud Storage does not switch to another key by itself.
BThey cannot be accessed until the key version is enabled again
Correct
While a key version is disabled, data encrypted with it cannot be accessed; re-enabling restores access.
CThey stay readable because Google keeps a plaintext copy
Incorrect
Google does not keep a plaintext copy to bypass your key.
DThey are deleted immediately and permanently
Incorrect
Disabling a key version does not delete data; it blocks access.
With CMEK, your key controls access to the data. Disable a key version only when you mean to cut off access.
An analyst wants to know how many bytes a query over a large, non-clustered table will process, and so roughly what it will cost on on-demand pricing, before running it. What should the analyst do?
ARun the query with LIMIT 10 to see the cost
BPreview the table in the console
CRun the query once and check the cost afterward
DPerform a dry run of the query
Show the answer and why
ARun the query with LIMIT 10 to see the cost
Incorrect
On non-clustered tables a LIMIT clause does not reduce the bytes processed or the cost.
BPreview the table in the console
Incorrect
Previewing shows rows of the table, not what a query would process.
CRun the query once and check the cost afterward
Incorrect
This pays for the query before knowing what it costs.
DPerform a dry run of the query
Correct
A dry run estimates the bytes the query would process without running it.
Estimate first with a dry run, and cap risk with a maximum bytes billed setting when needed.