PCD Designing highly scalable, available, and reliable cloud-native applications Practice Test 11 — 15 Questions

Question 1

A financial services company runs a transaction processing microservice on Google Kubernetes Engine (GKE). The service uses Cloud Spanner as its database. After migrating from Cloud SQL to Spanner to improve scalability, the team notices that a small percentage of transactions fail with an 'ABORTED' error due to deadlock detection. The application currently performs no retries, and the failures cause customer-facing errors. The team also observes that under peak load, transaction latencies are around 500ms, which is acceptable but they want to ensure the system remains reliable. They need to implement a solution that minimizes failures while maintaining acceptable performance. Which course of action should they take?

Choose an answer to begin — your selection is scored in the full session.

Accepted Answer

Implement retry logic with exponential backoff and random jitter for aborted transactions.. In Cloud Spanner, 'ABORTED' errors due to deadlock detection are a normal part of the optimistic concurrency control mechanism. The correct solution is to implement retry logic with exponential backoff and random jitter, as recommended by Google's own documentation. This approach transparently handles transient deadlocks without requiring infrastructure changes or sacrificing consistency, and it maintains acceptable latency by spacing out retries.

Answer

Increase the number of Spanner nodes to reduce the probability of deadlocks.

Answer

Reduce the size of each transaction by splitting them into smaller ones.

Answer

Change the transaction isolation level to READ UNCOMMITTED to avoid deadlocks.

Question 2

A retail company runs an order-processing service on Cloud Run (fully managed) in us-central1. During flash sales, order volume spikes 20x for about 10 minutes. The service must remain responsive, minimize cold starts, and keep costs predictable. The team wants to configure the service to handle these bursts. What should they do?

Accepted Answer

Set the minimum number of instances to a value that covers baseline traffic and configure a maximum instance limit, along with concurrency settings appropriate for the service.. Cloud Run's autoscaling can handle bursts, but to minimize cold starts and keep costs predictable, you should set a minimum instance count to keep some instances warm and a maximum instance limit to cap scaling. Tuning concurrency ensures efficient request handling per instance. This balances responsiveness and cost control for flash-sale traffic.

Answer

Configure the service to use a larger container image and increase the memory limit to reduce cold-start time.

Answer

Deploy the service to a GKE cluster with cluster autoscaler and horizontal pod autoscaler, and route traffic through an HTTP(S) load balancer.

Answer

Enable CPU always allocated and set the maximum number of instances to a high value, without setting a minimum.

Question 3

A financial services company runs a portfolio management application on GKE Autopilot. The application must remain available even if an entire zone in a region goes offline. The team wants to minimize manual intervention and ensure that pods are automatically rescheduled when a zone fails. Which configuration should they implement?

Accepted Answer

Create a regional GKE cluster with a node pool spanning multiple zones and use topology spread constraints to distribute pods evenly across zones.. A regional GKE cluster with a multi-zone node pool and topology spread constraints ensures that pods are evenly distributed across zones. If one zone fails, the remaining pods continue to serve traffic, and Kubernetes automatically reschedules pods onto healthy nodes. This design provides high availability without manual intervention.

Answer

Deploy the application on a single-zone GKE cluster with high-availability control plane and use a Horizontal Pod Autoscaler to scale replicas during zone outages.

Answer

Configure a node pool with nodes spread across multiple zones and set pod anti-affinity rules to avoid scheduling multiple replicas on the same node.

Answer

Use a managed instance group of Compute Engine instances with an autoscaler and deploy the application directly on VMs, then configure a global load balancer.

Question 4

A financial services company runs a critical transaction processing system on Compute Engine. The system uses a regional managed instance group (MIG) with autohealing and autoscaling. They need to ensure that the application remains available even if an entire zone fails, and that recovery is automatic. They also want to minimize downtime during planned maintenance. What should they do?

Accepted Answer

Deploy the application across multiple zones using a regional MIG with autohealing, and use a global load balancer with health checks to route traffic only to healthy instances.. A regional MIG spreads instances across multiple zones, so a zone outage only impacts a fraction of capacity. Autohealing replaces unhealthy instances, and a global load balancer with health checks routes traffic to healthy instances. This design provides automatic recovery from zone failures and reduces downtime during maintenance by allowing rolling updates without service interruption.

Answer

Configure the MIG to span multiple zones within the region, and set the instance template to use a regional persistent disk for data.

Answer

Configure the MIG to use a stateful configuration with persistent disks, and set the update policy to opportunistic.

Answer

Use a zonal MIG and manually create a second MIG in another zone, then use DNS failover to switch traffic.

Question 5

A team runs a fault-tolerant web application on a regional managed instance group of Compute Engine VMs behind an external Application Load Balancer. They need to ensure that if an entire zone becomes unavailable, traffic continues to be served with minimal disruption and no manual intervention. Which configuration should they implement?

Accepted Answer

Configure the managed instance group to spread instances across multiple zones within the region and set the load balancer's health check to use a short interval.. A regional managed instance group spreads instances across multiple zones, so a zone outage only affects a fraction of capacity. The load balancer health check ensures traffic is routed only to healthy instances. This design provides automatic zone-level failover without manual intervention, which is exactly what the scenario requires. Cross-region failover is overkill for a single-zone failure.

Answer

Create a second managed instance group in a different region and use a global load balancer with failover based on a health check.

Answer

Use a single-zone managed instance group and enable autoscaling to add instances when a zone fails.

Answer

Deploy a standalone VM in each zone and manually add them to the load balancer backend service.

Question 6

A company runs a stateless API on Cloud Run that receives a steady stream of requests. They want to minimize cold starts and ensure predictable latency during traffic spikes. The service is containerized and uses a small image. Which configuration should they choose?

Accepted Answer

Set the minimum number of instances to a value greater than zero and enable CPU always allocated.. To minimize cold starts, you must keep instances warm by setting a minimum instance count greater than zero. Additionally, enabling CPU always allocated ensures that the instance has CPU available even when idle, which is required for background work and reduces latency. The other options either do not prevent cold starts or may worsen performance.

Answer

Increase the maximum number of instances and set the concurrency to 1.

Answer

Use a larger container image with more dependencies to reduce startup time.

Answer

Deploy the service with a higher memory limit and set the timeout to 15 minutes.

Question 7

A media streaming company is designing a new application on Google Cloud. The application must be highly available across regions and able to survive a full region outage. They want to minimize recovery time and data loss. Which TWO strategies should they implement? (Choose two.)

Accepted Answer

Deploy the application in multiple regions and use a global external HTTP(S) load balancer to route users to the closest healthy region.. Deploying across multiple regions with a global load balancer ensures traffic is routed to a healthy region if one fails. Using dual-region or multi-region Cloud Storage with Cloud CDN ensures media assets are replicated and served with low latency. Together, these strategies provide cross-region high availability and fast recovery with minimal data loss.

Answer

Configure Cloud SQL for MySQL with a read replica in a different region and promote it manually during a regional failure.

Answer

Deploy the application on a single Compute Engine instance with a global static IP address and take regular snapshots.

Answer

Use a single regional managed instance group with autoscaling and a regional persistent disk for the application state.

Question 8

A team runs a Cloud Run service that occasionally experiences cold starts, causing latency spikes. They want to ensure that at least one instance is always ready to handle requests without cold start delays, while minimizing cost. What should they do?

Accepted Answer

Set the minimum number of instances to 1.. Maintaining a minimum instance count of 1 ensures that at least one instance remains warm and ready to serve requests, directly addressing cold start latency. The other options affect scalability limits, startup speed, or request handling duration, but none guarantee a running instance during idle periods.

Answer

Enable CPU boost during startup.

Answer

Increase the maximum number of instances.

Answer

Configure a longer request timeout.

Question 9

A startup is building a new mobile backend using Cloud Run. They expect unpredictable traffic and want to minimize costs while ensuring the service scales automatically. What should they do?

Accepted Answer

Deploy the service to Cloud Run with the default settings, which automatically scale based on incoming requests.. Cloud Run is a serverless platform that automatically scales based on request load and scales to zero when idle, making it cost-effective for unpredictable traffic. Default settings handle scaling without additional configuration, so the startup can focus on development.

Answer

Deploy the service to Compute Engine instances behind a load balancer and use an autoscaler based on CPU usage.

Answer

Deploy the service to Cloud Functions with a maximum instance limit set to 100 to control costs.

Answer

Deploy the service to a GKE cluster with a Horizontal Pod Autoscaler configured to scale based on CPU utilization.

Question 10

A retail company runs an order-processing service on Cloud Run (fully managed) in the us-central1 region. During flash sales, request latency spikes and some requests return HTTP 503 errors because all container instances are busy. The team wants the service to absorb sudden traffic bursts without manual intervention, while keeping cold-start latency low and controlling cost. Which combination of Cloud Run settings should they configure?

Accepted Answer

Set a minimum instance count greater than zero and increase the maximum instance count, relying on concurrency and CPU allocation settings to handle bursts.. The service fails because concurrent requests exceed available container instances. Raising the maximum instance count allows horizontal scale-out, and a non-zero minimum instance count keeps warm instances ready so bursts do not trigger cold starts. Together they address both the 503 errors and latency. Platform changes or timeout and concurrency tweaks do not add the capacity needed for sudden traffic spikes.

Answer

Increase the Cloud Run request timeout to 60 minutes and lower the concurrency setting to 1 so each request gets a dedicated instance.

Answer

Enable Cloud CDN in front of the Cloud Run service and configure a global external HTTP(S) load balancer with Cloud Armor rate limiting.

Answer

Deploy the service to a GKE Autopilot cluster and configure a HorizontalPodAutoscaler based on CPU utilization.

Question 11

A logistics company runs a shipment tracking application on GKE. The application consists of microservices that communicate via gRPC. They need to ensure that if one microservice becomes slow or unresponsive, it does not cause cascading failures. They also want to limit the number of concurrent requests to each service and automatically retry failed requests with backoff. What should they do?

Accepted Answer

Use Istio service mesh with Envoy sidecar proxies to enforce circuit breaking, retries, timeouts, and connection pool limits.. Istio service mesh with Envoy sidecars provides centralized traffic management, including circuit breaking, retries with backoff, timeouts, and connection pool limits. These features prevent a slow or unresponsive microservice from causing cascading failures by limiting concurrent requests and isolating failures. This is more consistent and manageable than client-side configurations.

Answer

Deploy the microservices on separate GKE clusters and use a global load balancer to distribute traffic.

Answer

Configure Kubernetes Horizontal Pod Autoscaler (HPA) and Pod Disruption Budgets (PDBs) for each microservice.

Answer

Implement client-side load balancing with gRPC and configure retries and timeouts in each client.

Question 12

A company runs a stateful application on Google Kubernetes Engine (GKE) using a StatefulSet with persistent volumes. They need to ensure that the application can survive a zone failure without data loss and maintain high availability. Which two strategies should they implement? (Choose two.)

Accepted Answer

Configure persistent volumes with a Regional Persistent Disk (Regional PD) storage class.. To survive a zone failure with a stateful application, you need both compute and storage redundancy. A regional GKE cluster spreads nodes across zones, and Regional Persistent Disks replicate data across zones. Together, they allow the application to continue running and preserve data if a zone becomes unavailable.

Answer

Enable horizontal pod autoscaling based on CPU utilization.

Answer

Use a single-zone GKE cluster with node auto-repair enabled.

Answer

Set the pod anti-affinity to ensure pods are scheduled on different nodes.

Question 13

A company runs a microservices application on GKE. They want to ensure that if the Redis cache used by one service becomes unavailable, the service degrades gracefully and continues to serve requests with reduced functionality. Which pattern should they implement?

Accepted Answer

Fallback pattern by implementing logic in the service to use a local cache or default values when Redis is unavailable.. To degrade gracefully when Redis is unavailable, the service should implement a fallback mechanism, such as using a local cache or returning default data. This ensures continued operation with reduced functionality, rather than complete failure.

Answer

Circuit breaker pattern using Istio to detect failures and redirect traffic to a fallback service.

Answer

Bulkhead pattern to isolate the Redis connection pool and prevent resource exhaustion.

Answer

Retry pattern with exponential backoff to handle transient Redis failures.

Question 14

A healthcare company is building a patient portal on Google Cloud. The application must be highly available and able to scale automatically. They expect unpredictable traffic patterns. The application uses a stateless frontend and a Cloud SQL for PostgreSQL database. They want to minimize operational overhead. What should they do?

Accepted Answer

Deploy the frontend on Cloud Run, and configure Cloud SQL with a regional instance and automatic storage increase.. Cloud Run is a fully managed platform that automatically scales stateless containers, eliminating infrastructure management. A regional Cloud SQL instance provides high availability with automatic failover, and automatic storage increase prevents storage issues. This pairing meets the requirements for high availability, automatic scaling, and minimal operational overhead.

Answer

Deploy the frontend on a regional managed instance group with autoscaling, and use Cloud SQL with a single zone instance.

Answer

Deploy the frontend on GKE Autopilot, and use Cloud SQL with a read replica in another region.

Answer

Deploy the frontend on Compute Engine instances in a zonal MIG, and use Cloud SQL with a regional instance.

Question 15

A company is deploying a new microservices application on Google Cloud. They want to ensure that if one service fails, it does not cause a cascading failure to other services. They plan to use Cloud Run for each microservice. Which design pattern should they implement?

Accepted Answer

Circuit breaker pattern. The circuit breaker pattern is specifically designed to prevent cascading failures by quickly failing requests to a service that is likely to fail, giving it time to recover. This isolates failures and maintains overall system stability, which is essential in a microservices architecture on Cloud Run.

Answer

Load balancing across multiple instances

Answer

Asynchronous messaging with Pub/Sub

Answer

Retry with exponential backoff