Prepare and test your skills

Prepare and test your skills

Free sample question with answers and detailed explanation below. The Associate Cloud Engineer practice bank has an additional 2,014 practice questions.Try 10 free realistic questions: instant score, no credit card, and no account required.
Your company runs a web application distributed across multiple Compute Engine instances in a managed instance group. During routine background tasks, individual virtual machines (VMs) experience brief, transient CPU spikes that cause alerting policies to trigger false positive notifications. You need to configure a Cloud Monitoring alert policy that evaluates the sustained workload of the entire fleet and triggers only when the collective CPU utilization across all VMs exceeds an 80% threshold over a 5-minute window.
Which alignment and aggregation method should you configure for the time-series metric?
The Rate of change aligner measures how rapidly a metric value changes per unit of time (derivative) rather than reporting the current magnitude of the metric. Combining this with a Max cross-series reducer tracks the single fastest rate of increase observed across any node in the fleet.
This method would be appropriate for Early Anomaly and Surge Detection, such as detecting a sudden flash crowd or runaway loop by alerting when CPU usage surges upward at an abnormal rate (e.g., rising faster than +30% CPU per minute).
In Google Cloud Monitoring, processing time-series data involves two essential stages: alignment (regularizing and smoothing points within an individual time series over an alignment period) and cross-series reduction/aggregation (combining multiple time series across distributed resources into a single meaningful stream). Using the Mean (Average) function for both stages provides a smoothed, representative indicator of fleet-wide utilization.

Alternative reducers such as Max or Sum either preserve single-node spikes or scale upward with instance counts, making them unsuitable for percentage-based fleet thresholding. Using Mean at both the alignment and reduction stages directly matches the goal of tracking sustained, collective CPU performance.
The Sum aggregation function adds all values together within the alignment window and across all time series streams. For gauge metrics such as CPU utilization percentages, summing combines percentage values across all running VMs into an aggregate sum.
This configuration would be correct if you were monitoring Total Request Count or Aggregate Ingress Bandwidth across the fleet, where summing discrete numerical events over 5 minutes gives the total volume processed by all servers combined.
The Max aligner captures the highest single metric sample within each alignment period for every individual time series. Leaving cross-series aggregation set to None preserves each VM's time series as an independent stream evaluated separately against the alert threshold.
This configuration would be appropriate if your objective were a strict Single-Node Failure Detection Policy, where exceeding 80% CPU on any individual instance for even a single minute signals a critical process hang that requires immediate automated remediation.
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