PCDE Optimising Service Performance Practice Test 2 — 15 Questions

Question 1

A company runs a GKE cluster for a web application. During peak traffic, the application experiences increased latency. The team has enabled the Horizontal Pod Autoscaler (HPA) based on CPU utilization, but the scaling is not fast enough. Which approach would improve the responsiveness of the HPA?

Accepted Answer

Decrease the HPA sync period from 15 seconds to 5 seconds.. Decreasing the HPA sync period makes the controller evaluate metrics more frequently, reducing the time to react to load changes. Increasing stabilization window delays scaling up, making it worse.

Answer

Decrease the CPU target utilization from 80% to 50%.

Answer

Increase the HPA sync period from 15 seconds to 60 seconds.

Answer

Increase the stabilization window from 5 minutes to 10 minutes.

Question 2

A company is using Cloud Run to deploy a service that processes background tasks. The service takes a few seconds to initialize, and users experience high latency on cold starts. How can the company eliminate cold starts for this service?

Accepted Answer

Set the minimum number of instances to a value based on the baseline traffic.. Setting a minimum number of instances ensures that at least that many instances are always running and ready to serve requests, eliminating cold starts.

Answer

Set the maximum number of instances to a higher value.

Answer

Use the gen1 execution environment.

Answer

Set the concurrency to 1.

Question 3

A company uses Cloud Spanner for a global application. They notice high write latency and occasional hotspotting on a table with a monotonically increasing integer primary key. Which schema design change would best prevent hotspotting while maintaining read performance?

Accepted Answer

Use bit-reversed indexes for the primary key.. Using a hash prefix of the primary key distributes writes across different splits, preventing hotspots. UUIDs also work, but bit-reverse is better for sequential keys. Interleaved tables do not help with hotspotting.

Answer

Use a UUID as the primary key.

Answer

Create an interleaved table with the primary key as the parent.

Answer

Add a secondary index on the primary key.

Question 4

A company is running a Cloud SQL for PostgreSQL database and wants to improve read performance for reporting queries. They have already optimized the queries but still see high CPU usage on the primary instance. What is the most cost-effective solution?

Accepted Answer

Add a read replica and direct reporting traffic to it.. Read replicas offload read traffic from the primary instance, reducing CPU usage without requiring a larger primary instance.

Answer

Enable connection pooling with PgBouncer on the primary.

Answer

Increase the memory of the primary instance.

Answer

Create a Cloud SQL Auth Proxy and use it for all connections.

Question 5

A company uses Pub/Sub to ingest events from multiple services. They notice a backlog of unacknowledged messages and want to increase the throughput of their subscriber. The subscriber is a single process running on Compute Engine. What is the most effective way to increase throughput?

Accepted Answer

Add more subscriber instances running in parallel.. Running multiple subscriber instances (parallel pull consumers) increases the overall throughput by allowing more messages to be pulled and processed concurrently.

Answer

Decrease the flow control max outstanding messages.

Answer

Increase the acknowledgement deadline to 10 minutes.

Answer

Enable ordering keys on the subscription.

Question 6

A company runs a batch data processing job on Compute Engine that is fault-tolerant. They want to reduce costs without affecting job completion time. The job can handle instance preemption gracefully. Which compute option should they use?

Accepted Answer

Use preemptible VMs.. Preemptible VMs are significantly cheaper and suitable for fault-tolerant batch jobs because they can be interrupted but the job can resume on new instances.

Answer

Use regular VMs with committed use discounts for 1 year.

Answer

Use GPU-accelerated instances.

Answer

Use sole-tenant nodes.

Question 7

A company has a GKE cluster with cluster autoscaler enabled. They notice that after a batch job completes, the cluster takes a long time to scale down, leaving idle nodes running and incurring costs. Which configuration change would reduce the scale-down delay?

Accepted Answer

Reduce the scale-down delay from 10 minutes to 2 minutes.. The cluster autoscaler has a default scale-down delay (e.g., 10 minutes for unneeded nodes). Reducing this delay causes nodes to be removed sooner after becoming idle.

Answer

Enable node auto-provisioning.

Answer

Increase the min node count to match the peak.

Answer

Decrease the max node count.

Question 8

A company is using Cloud Run for a service that performs background computation even when there are no incoming requests. They find that the service is being throttled and not completing the background work. What is the most likely cause and solution?

Accepted Answer

The service must have CPU always-on enabled.. Cloud Run instances have CPU throttled when not handling requests. Setting CPU always-on ensures the instance can use CPU continuously for background tasks.

Answer

The service needs more memory; increase memory limit.

Answer

The service needs to set min instances to 1.

Answer

The service should be migrated to GKE.

Question 9

A team is using Cloud Spanner and wants to create a secondary index that is stored with the base table data to avoid interleaved joins. Which option should they use when creating the index?

Accepted Answer

Add `spanner_interleave_in_parent = true` in the index creation.. The `spanner_interleave_in_parent` option stores the index interleaved within the parent table, improving join performance for parent-child relationships.

Answer

Create a stored generated column as the index key.

Answer

Use a hash index instead of a secondary index.

Answer

Use the `INTERLEAVE IN` clause when creating the index.

Question 10

A company uses Cloud SQL for MySQL and notices slow queries. They have enabled slow query logging and found that some queries are performing full table scans. The table has millions of rows. What is the best immediate action to improve query performance?

Accepted Answer

Create an index on the columns used in the WHERE clause.. Creating an appropriate index can eliminate full table scans and drastically improve query performance for the identified slow queries.

Answer

Increase the instance memory to allow more caching.

Answer

Use `EXPLAIN ANALYZE` to rewrite the query.

Answer

Add a read replica to offload the queries.

Question 11

A company is deploying a microservice on Cloud Run that needs to handle up to 1000 concurrent requests per instance. The default concurrency setting is 80. How should they configure the service to achieve the desired concurrency?

Accepted Answer

Set the `--concurrency` flag to 1000.. Cloud Run allows setting max concurrent requests per instance via the `--concurrency` flag. Setting it to 1000 enables the desired throughput.

Answer

Set the `--execution-environment` to gen2.

Answer

Set the `--cpu` flag to 4.

Answer

Set the `--max-instances` flag to 1000.

Question 12

A company wants to use the Vertical Pod Autoscaler (VPA) to automatically adjust resource requests for their pods. They want the VPA to update the resource requests of running pods without recreating them. Which VPA updateMode should they use?

Accepted Answer

Auto. The correct mode is `Auto`. VPA in `Auto` mode automatically adjusts resource requests and limits by recreating pods. While the requirement states 'without recreating them', VPA's `Auto` mode does recreate pods because resource requests are immutable. However, `Auto` is the only mode that automatically applies recommendations; `Recreate` only applies on pod creation, `Initial` only at pod start, and `Off` provides recommendations only. Thus, `Auto` is the best choice for ongoing automatic adjustment.

Answer

Initial

Answer

Recreate

Answer

Off

Question 13

A company runs a GKE cluster and wants to ensure that during a planned node upgrade, their application remains available with minimal disruption. Which TWO steps should they take? (Choose two.)

Accepted Answer

Deploy multiple replicas of the application across different nodes.. PodDisruptionBudgets (PDBs) ensure a minimum number of pods are available during voluntary disruptions like upgrades. Using multiple replicas across nodes provides redundancy. Increasing max surge helps but is not a direct disruption mitigation.

Answer

Increase the max surge of the deployment to 100%.

Answer

Set the cluster autoscaler scale-down delay to 0.

Answer

Use a single replica per node to simplify management.

Question 14

A company is experiencing slow query performance on Cloud Spanner. They have identified a query that joins a parent table with a child table frequently. Which THREE design choices can improve the performance of this join? (Choose three.)

Accepted Answer

Use interleaved tables to store child rows within the parent row.. Interleaved tables physically store child rows with the parent, reducing cross-node communication. Secondary indexes help with non-key lookups. Using `spanner_interleave_in_parent` on indexes stores index entries with the parent. Splitting tables into smaller tables does not help joins. Bit-reversed keys prevent hotspots but not join performance.

Answer

Split the child table into multiple smaller tables.

Answer

Create a secondary index on the parent table's primary key.

Question 15

A company uses Pub/Sub for event-driven processing and wants to ensure exactly-once delivery for critical messages. Which TWO configurations are required? (Choose two.)

Accepted Answer

Enable message ordering on the subscription.. For exactly-once delivery, the subscription must have message ordering enabled and the subscriber must be idempotent. Pub/Sub also requires enabling exactly-once delivery flag (now GA).

Answer

Use multiple subscriber instances to increase throughput.

Answer

Set the acknowledgement deadline to 600 seconds.

Answer

Disable flow control to allow unlimited messages.