Planning and implementing a cloud solution

A startup is deploying a stateless Python web application that receives highly variable traffic, including long idle periods overnight. The team has no operations staff and wants a fully managed platform that automatically scales down to zero instances when there is no traffic to minimize cost, while requiring the least configuration effort. Which compute option should they choose?

Your data analytics team needs to run ad-hoc SQL queries across several petabytes of historical clickstream data stored as structured tables. Queries run a few times a day, must return aggregated results within minutes, and the team does not want to manage any servers or provision capacity in advance. Which Google Cloud data service should you choose?

A financial analytics company is building a platform that ingests millions of stock trade events per second and must support single-digit millisecond read latency for key-based lookups. The access pattern is always by a row key (symbol + timestamp), and the total dataset is expected to grow into the hundreds of terabytes. The team does not need SQL joins or ACID multi-row transactions. Which Google Cloud data service should they choose?

A manufacturing company is deploying thousands of IoT sensors that will emit device metrics at very high throughput (hundreds of thousands of writes per second). The team needs a fully managed database optimized for time-series and high-volume single-key lookups, with low-latency reads and writes, that can scale horizontally. Which Google Cloud data service should they choose?

Your company runs several internal microservices on Compute Engine VMs within a single VPC network. Developers need to reach these services using friendly hostnames like api.internal.example.com instead of internal IP addresses, and these names must NOT be resolvable from the public internet. What should you configure?

Your team runs a photo-sharing app. Whenever a user uploads an original image to a Cloud Storage bucket, you must automatically generate a thumbnail and write it to a second bucket. The processing code takes a few seconds per image, there is no steady traffic pattern, and the team wants to avoid managing any servers or clusters. Which compute option best fits this requirement?

Your team has built a stateless REST API packaged as a Docker container. Traffic is unpredictable and often drops to zero at night. You want a fully managed platform that scales automatically, charges only when requests are being served, and requires no cluster management. Which service should you deploy to?

Your team has a Python script that processes a daily CSV export from a partner. The job runs for about 8 minutes, needs 2 GB of memory, and must execute automatically once every night at 02:00. There is no incoming HTTP traffic and no need for the code to run at any other time. You want the most cost-effective managed option with minimal infrastructure to maintain. What should you do?

A retail company is building a new order-management application that requires strong transactional consistency, complex SQL joins across orders, customers, and inventory tables, and full ACID compliance. The team wants a fully managed relational database and expects a moderate dataset well under 10 TB. Which Google Cloud data service should they choose?

Your team runs a transactional web application backed by a Cloud SQL for PostgreSQL instance. The analytics team runs heavy read-only reporting queries that are slowing down the primary database and affecting customer-facing transactions. You need to offload these reporting queries without changing the application's write path or migrating to a different database engine. What should you do?

Your team stores application log files in a Cloud Storage bucket. The logs are frequently accessed for the first 30 days for troubleshooting, rarely accessed between 30 and 365 days for occasional audits, and must be retained for 7 years for compliance but are almost never read after the first year. You want to minimize storage costs while keeping the data in the same bucket. What should you do?

Your company hosts a media catalog whose image assets are served to a web application with users across North America, Europe, and Asia. The assets are accessed constantly and read latency for the global audience must be minimized while ensuring high availability. Which Cloud Storage bucket configuration should you choose when creating the bucket?

Your company keeps compliance documents in Cloud Storage. Analysts access each document roughly once a month, and the data must remain immediately available with millisecond retrieval when requested. You want to minimize storage cost without introducing retrieval latency for these occasional reads. Which storage class should you choose?

Your company stores private product manuals in a Cloud Storage bucket that has uniform bucket-level access enabled and no public access. A partner needs to download one specific PDF file directly, but the partner does not have a Google account and you must not grant any long-term IAM permissions. The download link should stop working after 2 hours. What should you do?

Your team needs to host a static marketing website (HTML, CSS, JavaScript, and images) that will be accessed frequently by users worldwide. The content is read constantly and rarely changes. You want the most cost-effective and operationally simple way to serve this content directly from Cloud Storage. Which storage class should you choose for the bucket?

Your team is deploying a new Cloud Storage bucket that will hold thousands of objects for an internal analytics application. Security requires that access be managed centrally and consistently using only IAM policies, and they explicitly want to prevent anyone from setting per-object ACLs that could accidentally expose individual files. What should you configure when creating the bucket?

Your team runs a memory-intensive in-memory caching application on Compute Engine. Benchmarking shows the workload needs 6 vCPUs but requires 60 GB of RAM to hold the working set. Standard predefined machine types either give too little memory (n2-standard-6 has 24 GB) or waste vCPUs and money (n2-highmem-8 has 8 vCPUs and 64 GB). You want to minimize cost while meeting both requirements. What should you do?

Your team runs a stateless order-processing service on Compute Engine. Traffic is steady during business hours but spikes unpredictably during promotions. You want the number of identical VMs to grow and shrink automatically based on demand, while ensuring instances that fail health checks are automatically recreated. Which configuration should you use?

Your team runs a nightly batch rendering job on Compute Engine. The job is fault-tolerant, can be restarted at any time, and does not need to complete within a strict deadline. Management wants to minimize compute costs as much as possible. Which VM configuration should you choose?

Your team runs a production application on a single Compute Engine VM with a 500 GB persistent disk. Management requires automated daily backups of the disk with 14-day retention, minimal ongoing operational effort, and the ability to restore quickly if the disk is corrupted. What should you do?