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.
Your team manages a latency-sensitive application backed by a Cloud Bigtable SSD cluster with autoscaling enabled. Reviewing Cloud Monitoring metrics over the past month reveals:
recommended_node_count_for_cpu fluctuates between 3 and 8 nodes.recommended_node_count_for_storage remains steady at 4 nodes.Current autoscaling settings are: 10 minimum nodes, 40 maximum nodes, 55% CPU target, and 2.5 TB (50%) storage target per node.
Which configuration change should you make to optimize costs while preserving performance and staying within autoscaling limits?
The CPU utilization target (configurable between 10% and 80%) determines how aggressively Cloud Bigtable adds compute nodes as workload CPU rises. Lowering this target instructs Bigtable to provision more nodes earlier so each node operates at a lower CPU percentage.
recommended_node_count_for_cpu upward, requiring roughly 57% more nodes to handle the exact same workload.The application were experiencing unacceptably high latency during traffic increases and needed Bigtable to scale up more aggressively to preserve response times.
Lowering the minimum number of nodes to 2 aims to allow Cloud Bigtable to scale down even further during periods of very low traffic or storage utilization, maximizing potential cost savings during idle windows.
recommended_node_count_for_storage is steadily 4 nodes. Because Bigtable always selects the highest node count recommended across CPU and storage dimensions, the cluster would not scale down to 2 nodes anyway without breaching the storage utilization target.You also reduced the maximum number of nodes to 20 or fewer (so 2 meets the 10% threshold) and your stored data shrank enough for recommended_node_count_for_storage to drop to 2 nodes or less.
In Cloud Bigtable, the minimum number of nodes defines the lower bound below which the autoscaler will never scale in a cluster. Bigtable continuously calculates recommended node counts for both CPU (recommended_node_count_for_cpu) and storage (recommended_node_count_for_storage), and provisions the highest of those two values—as long as it falls between the configured minimum and maximum limits.
40 x 10% = 4).Unlike modifying CPU or storage targets—which cannot overcome a high minimum floor—lowering the minimum node count to 4 directly addresses the root cause of over-provisioning while keeping peak capacity (40 nodes) and latency safeguards unchanged.
The storage utilization target specifies how many terabytes (or GiB) of compressed data each Cloud Bigtable node can hold before autoscaling adds more nodes. On an SSD cluster (which has a 5 TB per-node limit), increasing the target from 2.5 TB (50%) to 4 TB (80%) allows each node to store more data before storage triggers a scale-up.
recommended_node_count_for_storage already at 4 nodes, storage is not what keeps the cluster at 10 nodes. Leaving the minimum at 10 guarantees the cluster remains over-provisioned regardless of the storage target.The cluster had a lower minimum node count, stored a large dataset where storage was the scaling bottleneck, and supported a latency-insensitive batch workload rather than a latency-sensitive application.
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