Prepare and test your skills

Prepare and test your skills

Free sample question with answers and detailed explanation below. The Professional Cloud Database Engineer practice bank has an additional 769 practice questions.Try 10 free realistic questions: instant score, no credit card, and no account required.
A healthcare company operates high-throughput Cloud Bigtable instances in production. An automated ingestion pipeline authenticates using a dedicated service account and performs thousands of read and write requests per second, while analysts and administrators query data and inspect metadata using individual user accounts.
Your auditing policy must satisfy these requirements:
Which Data Access audit log configuration should you implement?
Cloud Audit Logs categorize Data Access operations into three configurable permission types: ADMIN_READ (reading metadata or configuration), DATA_READ (reading user-provided data), and DATA_WRITE (writing user-provided data). Because Data Access audit logs are disabled by default for most Google Cloud services, they must be explicitly enabled in the IAM audit configuration (AuditConfig). Within this configuration, you can specify exempted principals (exemptedMembers) so that selected identities do not generate Data Access audit log entries.
ADMIN_READ, DATA_READ, and DATA_WRITE for the Bigtable service ensures full visibility into metadata inspections, data queries, and data mutations performed by human analysts and administrators.ADMIN_READ, DATA_READ, and DATA_WRITE visibility to match compliance mandates.Unlike downstream exclusion filters, which discard logs only after Bigtable has already spent resources constructing them, exempted principals in the audit configuration balance granular human query visibility with minimal cost and zero extra serving overhead.
In Cloud Audit Logs, ADMIN_READ is a Data Access audit log category that records API calls reading resource metadata or configuration settings. Admin Activity audit logs, which are always enabled and cannot be disabled, record ADMIN_WRITE operations—user-driven API calls that create, modify, or delete resource metadata or configurations (such as creating a table or updating IAM permissions).
Neither ADMIN_READ logs nor Admin Activity audit logs record access to user-provided data stored inside the database. When analysts query rows or mutate cell values in Bigtable, those actions check DATA_READ and DATA_WRITE IAM permissions, not ADMIN_READ or ADMIN_WRITE. Enabling only ADMIN_READ leaves a critical compliance gap because user-initiated data reads and data writes go completely unrecorded.
ADMIN_READ tracks who viewed configurations or schema metadata, not who queried sensitive healthcare records.ADMIN_WRITE) logs track administrative changes, not row-level inserts, updates, or deletes.DATA_READ and DATA_WRITE alongside exempted principals, you cannot selectively audit human data queries while omitting automated pipelines.The compliance policy only required monitoring administrative configuration changes and metadata reads—such as listing tables or viewing IAM policies—without auditing queries against actual database records.
Google Cloud allows configuring Data Access audit logs at the organization, folder, or project level in the resource hierarchy. When an audit policy is defined at the organization level, it automatically applies to every existing and newly created folder and project beneath that organization, providing centralized compliance governance.
This approach fails for two reasons. First, audit log configurations are strictly inherited downward: you cannot disable a Data Access audit log at the project level if it was enabled in a parent organization or folder. Consequently, attempting to turn off DATA_READ and DATA_WRITE in the Bigtable project will fail or have no effect, leaving the high-volume pipeline fully logged. Second, even if project-level disabling were possible, turning off DATA_READ and DATA_WRITE for the entire project would stop auditing human analysts' data queries and writes, violating the compliance requirement. Enabling all three log types for all services organization-wide also inflates log volume across unrelated workloads.
DATA_READ and DATA_WRITE at the service level affects both automated service accounts and human analysts.Your organization mandated an unchangeable baseline of Data Access auditing across every project and service, and you wanted to prevent project administrators from disabling those logs locally.
Cloud Logging exclusion filters are rules applied to a log sink—such as the predefined _Default sink—that inspect incoming log entries and drop matching records before they are ingested into a log bucket. Excluding logs from ingestion prevents those entries from counting toward Cloud Logging ingestion and storage costs.
While an exclusion filter reduces billing in Cloud Logging, it operates too late in the logging pipeline to protect database performance. With an exclusion filter, Bigtable still generates and emits every DATA_READ and DATA_WRITE audit log entry for the pipeline service account before Cloud Logging discards it. Under a workload of thousands of requests per second, constructing and emitting those log entries imposes a continuous processing load on the Bigtable service components that serve your production data, violating the requirement to avoid extra database overhead.
_Default bucket, but does not eliminate upstream emission work.NOT operators compared to inclusion filters) to avoid accidentally dropping critical admin or human user events.The requirement was strictly to reduce Cloud Logging storage and ingestion costs for logs from a low-throughput service where upstream log-generation overhead is negligible, or if you needed to suppress Policy Denied audit logs, which cannot be disabled through IAM audit configuration.
Answer them yourself and get your score the moment you finish. No account needed.
Full access includes
Keep the momentum going with these hand-picked practice scenarios
Last updated Validated