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.
You manage a multi-database environment on Google Cloud that supports an application using Firestore in Native mode and Cloud Bigtable. You need to configure Cloud Monitoring alerting policies to detect active performance degradation and resource exhaustion while avoiding noisy alerts from brief spikes:
How should you configure these alerting policies in Cloud Monitoring?
A metric-threshold alerting policy in Cloud Monitoring evaluates numerical time-series data against a static or aggregated boundary over a specified duration. An alert is triggered only when the monitored metric stays above or below the configured threshold for the entire retest window (duration window), ensuring notifications reflect sustained issues rather than momentary fluctuations.
serviceruntime.googleapis.com/api/request_latencies metric on the Consumed API resource with a filter of resource.service == 'firestore.googleapis.com' isolates Native mode database traffic. Aggregating to the 99th percentile (p99) over a 5-minute rolling window with an upper threshold of 250 ms ensures alerts fire only when tail latency remains degraded across the full 5-minute period.bigtable.googleapis.com/cluster/storage_utilization metric on the Cloud Bigtable Cluster resource and filtering by cluster ID tracks capacity per cluster. Setting an Above threshold value of 70 with a 10-minute retest window detects sustained high disk consumption while filtering out brief compaction spikes.Unlike forecasting, log-scanning, or unaggregated policies, properly aligned metric-threshold conditions directly measure real-time database health and capacity using standard service metrics, rolling aggregations, and sustained duration windows.
SQL-based alerting policies in Observability Analytics run scheduled BigQuery SQL queries against log views stored in upgraded Cloud Logging buckets (via a linked BigQuery dataset). They trigger an alert when the query result meets either a row count threshold or a boolean condition across aggregated log entries.
The requirement was to analyze aggregate patterns across log entries—such as alerting when more than 25% of database audit or error logs over the last 60 minutes have a severity of ERROR.
Evaluating raw, unaggregated time series with a 0-second retest window (most recent value) and resource metadata labels instructs Cloud Monitoring to check every individual label combination on every single reporting interval without smoothing, grouping, or waiting for sustained violations.
most recent value) triggers an alert on any single data point that spikes above 250 ms or 70%, completely defeating the requirement to alert only after sustained 5-minute and 10-minute breaches.There is no valid production scenario for combining raw unaggregated data with metadata label filters in sensitive database alerts; instead, best practices require aggregating to the target resource level, filtering by monitored-resource labels, and applying appropriate retest windows.
A forecasted metric-value condition analyzes historical trends in a time series to predict whether a metric is likely to cross a specified threshold within a future forecast window. Rather than reacting to a threshold that has already been breached, it aims to give teams advance warning of gradual capacity growth.
The goal were strictly proactive capacity planning—for example, predicting whether a steadily growing database disk will reach 85% utilization within the next 24 hours so administrators can provision additional nodes before an outage occurs.
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