PCD Designing highly scalable, available, and reliable cloud-native applications Practice Test 4 — 15 Questions

Question 1

An organization runs a critical application on Compute Engine with a regional managed instance group. They want to achieve 99.99% availability. Which architecture should they use?

Accepted Answer

Regional MIG with instances in three zones. To achieve 99.99% availability, the architecture must tolerate both a zonal failure and a single instance failure. A regional managed instance group (MIG) with instances in three zones ensures that even if one zone becomes unavailable, the remaining two zones can still serve traffic, meeting the 99.99% uptime target. Three zones provide the necessary redundancy because a two-zone regional MIG can only survive a single zone failure but not a simultaneous instance failure in the remaining zone, whereas three zones allow for a rolling update or failure of one zone while still maintaining quorum.

Answer

Regional MIG with instances in two zones

Answer

Single zone MIG with multiple instances

Answer

Multi-region deployment with global load balancer

Question 2

A company uses Cloud Spanner for a financial application. They need to ensure strong global consistency but also minimize latency for writes. What schema design should they use?

Accepted Answer

Use parent-child table relationships with interleaved tables. Interleaved tables in Cloud Spanner physically co-locate parent and child rows on the same split, reducing cross-node coordination for strongly consistent reads and writes. This minimizes write latency by ensuring that related data is stored together, avoiding distributed transaction overhead for operations that span parent-child relationships.

Answer

Use secondary indexes

Answer

Use commit timestamps

Answer

Use a single table with interleaved indexes

Question 3

A developer is building an event-driven system using Cloud Pub/Sub. They need to ensure reliable message delivery and processing. Which three practices should they follow?

Accepted Answer

Use pull subscriptions with synchronous acknowledgment.. Pull subscriptions with synchronous acknowledgment (option B) ensure that a message is not acknowledged until the subscriber has successfully processed it. This prevents premature acknowledgment and message loss, because Cloud Pub/Sub will redeliver the message if the acknowledgment deadline expires without a synchronous ack. This is a core pattern for reliable processing.

Answer

Set a minimum number of delivery attempts.

Answer

Use message ordering.

Question 4

A company runs a microservices application on Google Kubernetes Engine. They use Cloud SQL for persistent data. Recently, during a traffic spike, the application experienced increased latency and some requests failed with timeout errors. The team observed that the Cloud SQL CPU utilization spiked to 100%, and the GKE pods had high memory usage. They are using a standard Cloud SQL tier (db-n1-standard-2). Which course of action would best improve the application's performance and reliability?

Accepted Answer

Implement caching with Memorystore for frequently accessed data.. The primary bottleneck is the Cloud SQL CPU spiking to 100% under heavy read traffic. Implementing Memorystore (Redis) caching offloads repeated read queries from the database, reducing CPU load and query latency. This directly addresses the root cause—database CPU exhaustion—without requiring a larger database instance or adding replicas that would still be limited by the same CPU.

Answer

Upgrade Cloud SQL to a higher tier with more CPU.

Answer

Increase the number of replicas in GKE to reduce load per pod.

Answer

Add read replicas to Cloud SQL.

Question 5

A company runs a containerized application on Google Kubernetes Engine (GKE) with a regional cluster. The application experiences intermittent slowdowns during peak hours. The team notices that the number of nodes is not scaling up quickly enough. The application consists of a frontend deployment with a HorizontalPodAutoscaler (HPA) targeting 80% CPU utilization, and the cluster has a Cluster Autoscaler enabled with a maximum of 10 nodes. During a recent spike, the HPA increased replicas, but the Cluster Autoscaler was slow to add nodes, causing the new pods to remain pending. What is the most likely cause of this delay?

Accepted Answer

The Cluster Autoscaler has a built-in delay before adding nodes to avoid flapping.. The Cluster Autoscaler includes a built-in cooldown period (default 10–15 minutes) to prevent flapping—rapidly adding and removing nodes in response to transient spikes. During this delay, pending pods cannot be scheduled on new nodes, which explains why the HPA increased replicas but the new pods remained pending. This is the most likely cause given that the cluster is regional and the autoscaler is enabled.

Answer

The cluster is configured with a single zone, limiting node pool expansion.

Answer

The HPA is using a custom metric that is not supported by the Cluster Autoscaler.

Answer

The node pool's autoscaling is limited by the quota for Compute Engine resources in that zone.

Question 6

A development team is deploying a new application on Cloud Run. They anticipate unpredictable traffic patterns and want to minimize cold start latency. They also need to ensure that the application can handle sudden spikes without request drops. Which configuration should they use?

Accepted Answer

Set min-instances to a non-zero value to keep some instances warm, and enable CPU always-on.. Setting min-instances to a non-zero value ensures that Cloud Run always keeps at least that many instances warm, eliminating cold starts for baseline traffic. Enabling CPU always-on prevents the instance's CPU from being throttled to zero when idle, allowing the instance to handle incoming requests immediately without a cold start penalty. This combination minimizes latency for unpredictable traffic and ensures capacity to absorb sudden spikes without dropping requests.

Answer

Use App Engine Standard Environment with automatic scaling.

Answer

Set min-instances to 0 and max-instances to a high number to allow scaling from zero.

Answer

Use Cloud Functions instead of Cloud Run for better cold start performance.

Question 7

A company runs a critical financial application on Google Cloud using Compute Engine instances in a managed instance group (MIG) with auto-scaling based on CPU utilization. The application stores state in a local SSD and relies on sticky sessions (session affinity). Recently, during a traffic spike, the MIG scaled out new instances, but some users lost their sessions because the load balancer routed them to a different instance. The team needs to maintain session persistence without sacrificing scalability. What should they do?

Accepted Answer

Implement a shared session store using Cloud Memorystore for Redis.. Using Cloud Memorystore for Redis provides a centralized, external session store that decouples session state from individual Compute Engine instances. This ensures that any instance in the managed instance group can serve any user's request, maintaining session persistence even as the MIG scales out or in based on CPU utilization. It preserves scalability because the session store is independent of instance lifecycle, and Redis offers low-latency reads and writes suitable for session data.

Answer

Increase the instance group's cooldown period to reduce scaling frequency.

Answer

Use a global HTTPS Load Balancer with cookie-based session affinity.

Answer

Use Cloud NAT for consistent source IP routing.

Question 8

A large e-commerce platform uses Cloud Bigtable to store user session data and product recommendations. They have a single cluster in a single zone. During a recent zone outage, the application became unavailable for 30 minutes because Cloud Bigtable was unreachable. The team needs to ensure high availability for the session data with a Recovery Time Objective (RTO) of less than 5 minutes and a Recovery Point Objective (RPO) of zero (no data loss). What should they do?

Accepted Answer

Add a second cluster in a different zone within the same region and use multi-cluster routing to automatically failover.. Adding a second Cloud Bigtable cluster in a different zone within the same region and enabling multi-cluster routing provides automatic failover with an RTO of under 5 minutes and an RPO of zero. Multi-cluster routing directs read and write requests to the nearest healthy cluster, and replication between clusters is synchronous within a region, ensuring no data loss during a zone outage.

Answer

Migrate the session data to Cloud Memorystore for Redis with persistence and replication.

Answer

Configure replication to a second cluster in a different region and use global routing to failover.

Answer

Use Cloud Bigtable's single-cluster replication to a different zone.

Question 9

A team is deploying a new Cloud Run service that calls a third-party API which occasionally returns HTTP 503 errors. The team wants the service to automatically retry these failed requests to improve reliability, but they do not want to modify the application code. What should they do?

Accepted Answer

Implement retry logic in the application code using a library such as Google API Core or an exponential backoff algorithm.. Cloud Run does not provide built-in retry mechanisms for outbound HTTP requests. To handle transient failures from third-party APIs, the application must implement its own retry logic, often using exponential backoff or a client library. This ensures reliability without relying on platform features that do not exist.

Answer

Configure the Cloud Run service with a retry policy using the gcloud run deploy command with the --retry flag.

Answer

Enable the Cloud Run service's automatic retry feature by setting the --max-retries parameter to 3.

Answer

Set the Cloud Run service's concurrency to 1 to ensure requests are processed sequentially and retried automatically.

Question 10

A company is deploying a stateful web application on Google Kubernetes Engine (GKE) that requires persistent storage. They need to ensure high availability and data durability across zones. Which TWO strategies should they implement? (Choose two.)

Accepted Answer

Configure pod anti-affinity rules to ensure that replicas are scheduled on nodes in different zones.. For a stateful application on GKE requiring high availability across zones, using regional persistent disks ensures storage durability and failover, while pod anti-affinity spreads replicas across zones to avoid a single point of failure. Together, they provide both compute and storage resilience against zone outages.

Answer

Set the application's PodDisruptionBudget to allow zero disruptions to ensure continuous availability.

Answer

Use a single zonal persistent disk and enable automatic backup to Cloud Storage.

Answer

Deploy the application as a StatefulSet with multiple replicas and use a shared Filestore instance for persistent storage.

Question 11

A media streaming company uses Cloud Run to serve video metadata to a global audience. They want to minimize latency for users in Europe, Asia, and North America while keeping operational costs low. The service is stateless and can be deployed in multiple regions. What should they do?

Accepted Answer

Deploy the Cloud Run service in multiple regions and use a global external HTTP(S) load balancer to route users to the nearest healthy region.. A multi-region Cloud Run deployment with a global external HTTP(S) load balancer provides low latency by serving users from the nearest region and ensures high availability through health-check-based failover. Cloud Run's scale-to-zero model keeps costs low when traffic is low.

Answer

Deploy the Cloud Run service in a single region and use a global external HTTP(S) load balancer with Cloud CDN to cache responses at edge locations.

Answer

Deploy the Cloud Run service in multiple regions and configure a DNS-based load balancing solution using Cloud DNS with geolocation routing policies.

Answer

Deploy the Cloud Run service in a single region and use Cloud Interconnect to connect to points of presence worldwide for faster access.

Question 12

A company is designing a new application that must be highly available and resilient to zonal failures. They plan to use Cloud Storage for storing user-uploaded files and Cloud SQL for relational data. They need to ensure that both services can survive a zone outage without data loss and with minimal downtime. Which two configurations should they implement? (Choose two.)

Accepted Answer

Use a Cloud Storage bucket with the Regional location type.. For Cloud SQL, a regional instance provides automatic failover to a standby in another zone, ensuring database availability during a zone outage. For Cloud Storage, a Regional bucket already replicates data across zones within the region, so it survives zone failures by default. Together, these configurations deliver the required high availability and data durability without unnecessary cross-region complexity.

Answer

Use a Cloud Storage bucket with the Multi-Region location type.

Answer

Configure Cloud SQL with a single-zone instance and enable automated backups.

Answer

Use a Cloud Storage bucket with the Standard storage class and enable Object Versioning.

Question 13

A financial analytics team stores time-series trade data in Cloud Bigtable. Their application performs millions of point reads per second using row keys formatted as #. They notice that reads for a small number of very active accounts are slower than reads for other accounts, and CPU utilization is uneven across nodes. They want to improve read performance and distribute load evenly without changing the application's read pattern. What should they do?

Accepted Answer

Pre-split the table using a row key design that distributes account data across multiple tablets, or use a key prefix that spreads hot accounts across nodes.. Bigtable performance depends heavily on row-key design, because each row range is served by a single tablet on a single node. Hot accounts concentrated in one key range create a hotspot that additional nodes cannot absorb. Pre-splitting the table and designing keys that spread hot accounts across tablet boundaries distributes load and improves read performance without changing how the application queries data.

Answer

Increase the number of nodes in the cluster and enable SSD storage instead of HDD storage.

Answer

Change the row key to a salted key that prepends a hash of the account_id, and update the application to compute the salt on reads and writes.

Answer

Enable a replication cluster in a second zone and configure the application to read from the nearest replica.

Question 14

A team runs a stateless container image on Cloud Run that queries a Cloud SQL for PostgreSQL instance. During a traffic spike, the service scales from 5 to 60 instances within one minute and many requests fail with 'connection refused' or 'too many clients already'. The team wants the service to keep scaling automatically without exhausting database connections. What should they do?

Accepted Answer

Connect through the Cloud SQL Auth Proxy using a Unix socket and enable a server-side connection pooler such as PgBouncer, sizing the pool to the database connection limit.. Cloud Run autoscaling multiplies the number of independent container instances, and each instance creates its own database connections, so a spike can exhaust Cloud SQL's max_connections. Routing traffic through the Cloud SQL Auth Proxy over the /cloudsql Unix socket and adding a server-side pooler like PgBouncer keeps the number of real PostgreSQL backends bounded while still allowing Cloud Run to scale out to many instances.

Answer

Increase the Cloud SQL instance tier and max_connections, then rely on the application's per-instance connection pool to handle the spike.

Answer

Deploy the Cloud SQL Auth Proxy as a sidecar container in the same Cloud Run revision and configure the application to connect to localhost.

Answer

Set the Cloud Run maximum instance count to a low value so the number of concurrent database connections stays within the Cloud SQL limit.

Question 15

A retail company uses Pub/Sub to process orders. The order-processing service is deployed on GKE and uses a single subscription with a pull-based subscriber. During peak hours, the service experiences high latency, and some messages are processed multiple times. The team wants to improve throughput and reduce duplicate processing. What should they do?

Accepted Answer

Enable exactly-once delivery on the subscription and ensure the subscriber uses the appropriate client library features.. Enabling exactly-once delivery on the Pub/Sub subscription ensures that each message is delivered and processed only once, eliminating duplicates. Combined with scaling the number of subscriber instances, this improves throughput while maintaining message integrity.

Answer

Create multiple subscriptions for the same topic and have each subscriber instance use a separate subscription.

Answer

Increase the acknowledgment deadline for the subscription and add more subscriber instances.

Answer

Switch to a push-based subscription and deploy a Cloud Function to process messages.