Google PCA Analysing and Optimising Technical and Business Processes — 15 Questions

Question 1

A company wants to reduce costs for a batch analytics job that runs nightly for 4 hours on Compute Engine VMs. The job is fault-tolerant and can handle instance restarts. Which Compute Engine VM pricing model is MOST cost-effective?

Accepted Answer

Preemptible VMs. Preemptible VMs offer the lowest cost (up to 80% discount) and are ideal for fault-tolerant, short-lived batch workloads that can handle interruptions. Sustained use discounts apply automatically but require running a VM for at least 25% of a month. Committed use discounts require a 1- or 3-year commitment and are not as flexible for a short nightly job.

Answer

3-year committed use discount (CUD)

Answer

1-year committed use discount (CUD)

Answer

Sustained use discounts

Question 2

A company is migrating an on-premises PostgreSQL database (5 TB) to Cloud SQL. They need minimal downtime and automated schema conversion if needed. Which GCP service should they use?

Accepted Answer

Database Migration Service (DMS). Database Migration Service (DMS) supports continuous migration with minimal downtime from on-premises PostgreSQL to Cloud SQL. It handles schema and data migration. Migrate for Compute Engine is for VM migration, Transfer Appliance is for bulk data transfer, and Datastream is for streaming replication to BigQuery.

Answer

Datastream

Answer

Migrate for Compute Engine (formerly Velostrata)

Answer

Transfer Appliance

Question 3

A company runs a web application on Compute Engine behind a Global HTTPS Load Balancer. Users report slow page loads, especially for static assets. The development team wants to cache content closer to users without modifying code. Which GCP service should they enable?

Accepted Answer

Cloud CDN. Cloud CDN uses Google's global edge cache to deliver content close to users. It can be enabled on the load balancer backend buckets (for Cloud Storage) or backend services (for Compute Engine). Cloud Armor is for security, Cloud DNS for domain resolution, and Cloud NAT for outbound connectivity.

Answer

Cloud NAT

Answer

Cloud Armor

Answer

Cloud DNS

Question 4

A company has a BigQuery dataset with a growing fact table (500 million rows, added daily). Queries that filter on a date column and group by a product ID are slow. The team wants to optimise query performance without increasing slot costs. Which two actions should they take? (Choose TWO.)

Accepted Answer

Cluster the table by the product ID column. Partitioning by date and clustering by product ID can significantly improve query performance: partitioning prunes scans to relevant days, and clustering co-locates rows with similar product IDs. Materialised views pre-aggregate data but increase storage costs. SELECT * is inefficient and should be avoided. Slots pricing does not help performance without changing configuration.

Answer

Switch to BigQuery Editions with Autoscaling slots

Answer

Use SELECT * only when necessary

Answer

Create a materialised view that pre-aggregates the data

Question 5

A company wants to reduce Cloud Storage costs for archival data that is accessed less than once a year. They have data stored in Standard storage class. Which TWO actions should they take? (Choose TWO.)

Accepted Answer

Use a lifecycle rule to delete objects that are older than 365 days. Using lifecycle policies to transition from Standard to Archive storage class reduces cost for rarely accessed data. Also, deleting unnecessary objects reduces storage. Coldline is for 90 days, Nearline for 30 days; Archive is for long-term archival. Changing object names does not reduce cost.

Answer

Set a lifecycle rule to transition objects older than 90 days to Coldline storage class

Answer

Rename objects to start with 'archive/' to reduce costs

Answer

Set a lifecycle rule to transition objects older than 30 days to Nearline storage class

Question 6

A company is designing a disaster recovery strategy for a critical application. They need to ensure that if the primary region fails, the application can fail over to a secondary region with a Recovery Point Objective (RPO) of 1 hour and a Recovery Time Objective (RTO) of 15 minutes. The application uses Cloud Spanner. Which Cloud Spanner configuration meets these requirements?

Accepted Answer

Multi-region configuration (e.g., nam3). Cloud Spanner multi-region configurations provide automatic replication across regions with synchronous replication, achieving RPO near zero and RTO in seconds. Single-region configurations do not provide geographic failover. A two-region setup like nam3 (US Central + US East) or eur3 (Europe) provides synchronous replication with strong consistency. Active/active is the default. Multi-region configs are recommended for DR.

Answer

Single-region configuration

Answer

Backup and restore in another region

Answer

Dual-region with asynchronous replication

Question 7

A company runs a Bigtable instance for real-time analytics. They notice increasing latency for point reads. The row key pattern is 'YYYYMMDD_userID_productID'. The team suspects read hotspots. Which row key design change would BEST distribute the load across tablets?

Accepted Answer

Reorder to 'userID_productID_YYYYMMDD'. Using a field prefix (like user ID) before the timestamp distributes reads across tablets because Bigtable orders rows lexicographically. If the timestamp is first, recent data falls into a single tablet, causing hotspots. Reversing the timestamp or hashing the row key are also common strategies. Salting with a random prefix is another approach.

Answer

Prepend a random number field: 'random_YYYYMMDD_userID_productID'

Answer

Keep the same order but use a shorter timestamp: 'YYMMDD_userID_productID'

Answer

Reverse the timestamp: 'DDMMYYYY_userID_productID'

Question 8

A company wants to monitor the performance of their application against a Service Level Objective (SLO) of 99.9% availability over a 30-day rolling window. They plan to use Google Cloud's operations suite (formerly Stackdriver). Which metric should they use as the Service Level Indicator (SLI) for availability?

Accepted Answer

Successful request ratio. Availability is typically measured as the fraction of successful requests (e.g., HTTP 200-499) vs total requests. The other options are valid SLIs for other aspects: latency, error rate, or saturation.

Answer

Error rate

Answer

Request latency

Answer

CPU utilization

Question 9

A company is migrating a 200 TB on-premises file server to Cloud Storage. The network bandwidth is limited to 100 Mbps. The migration must complete within 30 days. Which approach should they use?

Accepted Answer

Use Transfer Appliance. Transfer Appliance is a physical device that can transfer up to 480 TB in a single shipment. At 100 Mbps, transferring 200 TB over the network would take approximately 200 TB * 1024 GB/TB * 8 bits/byte / (100 Mbps) = 1,638,400 Gb / 0.1 Gbps = 16,384,000 seconds ≈ 190 days, far exceeding 30 days. Therefore, using Transfer Appliance is the only feasible option.

Answer

Use Storage Transfer Service from another cloud

Answer

Use gsutil rsync over the network

Answer

Use Cloud Data Fusion

Question 10

A company wants to optimise costs for their BigQuery workloads. They have a predictable set of queries that run every hour, with occasional ad-hoc queries from analysts. They want to reduce the cost of the scheduled queries while keeping the ad-hoc queries cost-effective. Which pricing model should they choose?

Accepted Answer

Use Editions with a baseline of reserved slots for scheduled queries and allow autoscaling for ad-hoc. BigQuery slot reservations (editions) allow allocating dedicated slots for predictable workloads, with separate pricing for ad-hoc queries (on-demand). Flat-rate pricing is cheaper for large steady-state usage but may waste slots during idle time. On-demand is pay-per-query; for predictable queries, it may be expensive. The best approach is to use a hybrid: reserved slots for scheduled queries and on-demand for ad-hoc.

Answer

On-demand pricing

Answer

Use committed use discounts on slots

Answer

Flat-rate pricing (slot reservations) for all queries

Question 11

A company running a globally distributed application uses Firestore in Native mode. They need strong consistency for reads and writes across multiple regions. Which deployment strategy should they use?

Accepted Answer

Use Cloud Spanner. Firestore supports multi-region replication but only provides eventual consistency across regions. Cloud Spanner offers strong consistency globally. Firestore in Datastore mode also offers eventual consistency. Therefore, for strong consistency across regions, Cloud Spanner is required.

Answer

Use Bigtable with replication

Answer

Use Firestore in Datastore mode

Answer

Use Firestore in multi-region (nam5 or eur3)

Question 12

A company is planning to migrate a large on-premises Oracle database (10 TB) to Cloud SQL for PostgreSQL. They need to minimise downtime and ensure data integrity. Which TWO services or tools should they use? (Choose TWO.)

Accepted Answer

Database Migration Service (DMS). Database Migration Service (DMS) supports online migration from Oracle to Cloud SQL for PostgreSQL with minimal downtime by using continuous replication. Cloud SQL Auth Proxy is a tool for secure client connections to Cloud SQL instances; it is not required for the migration process itself. Cloud Dataflow is for data processing, Cloud Scheduler for job scheduling, and Migrate for Compute Engine for VM migration.

Answer

Migrate for Compute Engine

Answer

Cloud Dataflow

Answer

Cloud Scheduler

Answer

Cloud SQL Auth Proxy

Question 13

A company uses Cloud CDN to serve static content from a Cloud Storage bucket. They update the content every 2 hours, but users sometimes see stale content for up to 24 hours. They need users to see the latest content within 5 minutes of an update. Which action should they take?

Accepted Answer

Set a short max-age TTL (e.g., 5 minutes) on the objects. To ensure content is served fresh, they should set a short max-age (cache TTL) on the object metadata or via the bucket's default object holding. Invalidating the cache is a reactive measure, but setting a short TTL ensures that cache entries expire quickly. Cache modes affect query string handling, not TTL. Cache keys affect how content is cached but not the freshness duration.

Answer

Use cache invalidation after each update

Answer

Use custom cache keys

Answer

Change cache mode to CACHE_ALL_STATIC

Question 14

A company runs a batch processing job that uses preemptible VMs. The job occasionally fails due to VM preemption. They want to improve reliability without significantly increasing cost. Which TWO actions should they take? (Choose TWO.)

Accepted Answer

Use a managed instance group with autoscaling and preemptible VMs. Using a managed instance group (MIG) with autoscaling and preemptible VMs can automatically recreate instances if preempted. Implementing a retry logic in the application to re-run failed tasks also improves reliability. Using sole-tenant nodes or adding more nodes increases cost or complexity. Switching to regular VMs increases cost.

Answer

Use sole-tenant nodes to reduce risk of preemption

Answer

Switch to regular (non-preemptible) VMs

Answer

Increase the number of preemptible VMs

Question 15

A company is adopting Site Reliability Engineering (SRE) practices. After a major incident, they want to conduct a review to understand what went wrong and how to prevent recurrence, without blaming individuals. Which SRE practice should they follow?

Accepted Answer

Conduct a blameless postmortem. A blameless postmortem focuses on learning from incidents without assigning blame. Error budgets are for measuring reliability, SLOs/SLIs are for defining targets, and capacity planning is for scaling.

Answer

Define SLOs and SLIs

Answer

Create an error budget policy

Answer

Perform capacity planning