▸case-01 We ran our GitHub Actions CI pipeline on commit e4f1a9 and noticed p99 latency spiked from 45ms to 180ms compared to main branch commit c8b2f1. Analyze the CI pipeline execution logs to determine if this build introduces a performance regression. | pass→fail | 9,232 | 23,989 | +160% | 1 | 1 | 0% | 1,368 | 2,094 | +53% | 0 | 0 | — |
▸case-02 Execute a 5-minute load test using k6 against https://api.staging.example.com/v1/checkout with 50 concurrent virtual users and generate the raw pprof CPU flamegraph from the Go runtime. | pass→pass | 17,192 | 9,242 | -46% | 1 | 1 | 0% | 2,967 | 1,715 | -42% | 0 | 0 | — |
▸case-03 Package the benchmark results from our latest release candidate into a QA test summary report with pass/fail criteria for the Quality Assurance sign-off meeting. | pass→pass | 16,771 | 3,592 | -79% | 1 | 1 | 0% | 2,857 | 659 | -77% | 0 | 0 | — |
▸case-04 Our PostgreSQL database CPU utilization spikes to 95% during peak hours. Slow query logs show `SELECT * FROM orders WHERE customer_id = $1 ORDER BY created_at DESC` taking 3.2 seconds on a table with 15 million rows. Devs suggest rewriting the query to use subqueries. How should we optimize this bottleneck? | fail→fail | 13,163 | 16,564 | +26% | 1 | 1 | 0% | 2,235 | 2,806 | +26% | 0 | 0 | — |
▸case-05 An API endpoint `/users/active/details` takes 4.5 seconds to return 100 users. APM traces show 1 SQL query to fetch users followed by 100 individual queries to fetch user preferences. A developer proposed upgrading the RDS instance class from db.m5.large to db.m5.2xlarge. What is the proper fix? | pass→pass | 11,143 | 12,057 | +8% | 1 | 1 | 0% | 1,850 | 2,180 | +18% | 0 | 0 | — |
▸case-06 Google Search Console reports our main product page fails Web Vitals with an Interaction to Next Paint (INP) of 620ms. The team wants to compress all PNG hero images to WebP to fix INP. What optimization steps should be prioritized? | fail→pass | 15,356 | 14,193 | -8% | 1 | 1 | 0% | 2,395 | 2,538 | +6% | 0 | 0 | — |
▸case-07 Our news site has a Cumulative Layout Shift (CLS) score of 0.42 because dynamically injected ads and unsized images render late. A developer suggested setting CSS `display: none` until all fonts and scripts load. How should this be resolved? | pass→pass | 16,063 | 16,130 | +0% | 1 | 1 | 0% | 2,652 | 2,553 | -4% | 0 | 0 | — |
▸case-08 Our homepage Largest Contentful Paint (LCP) takes 4.8 seconds. Analytics show three synchronous external CSS files and two custom web fonts blocking initial rendering in the HTML `<head>`. What is the recommended strategy? | fail→fail | 13,544 | 12,084 | -11% | 1 | 1 | 0% | 2,461 | 2,329 | -5% | 0 | 0 | — |
▸case-09 During high traffic spikes, our Express.js backend throws `Error: Connection pool exhausted` and response times jump from 50ms to 8000ms. The DevOps lead suggests doubling backend pod counts in Kubernetes. What is the root optimization? | pass→pass | 12,889 | 14,183 | +10% | 1 | 1 | 0% | 2,090 | 2,450 | +17% | 0 | 0 | — |
▸case-10 A microservice handles 20,000 HTTP requests per second for static product catalog metadata. Every request executes a complex SQL join query directly on PostgreSQL. What architectural change should be prioritized? | fail→pass | 12,013 | 11,315 | -6% | 1 | 1 | 0% | 1,773 | 1,864 | +5% | 0 | 0 | — |
▸case-11 Our AWS monthly bill shows $4,200 in NAT Gateway data processing fees because S3 backups from EC2 instances inside private subnets flow through the NAT Gateway. The finance team wants us to downsize instance sizes. How do we reduce this infrastructure cost? | fail→fail | 12,288 | 11,071 | -10% | 1 | 1 | 0% | 1,990 | 2,139 | +7% | 0 | 0 | — |
▸case-12 CloudFront CDN billing shows a 12% cache hit ratio for static images because the origin sends `Cache-Control: private, max-age=0`. Engineers want to add 5 origin server instances to handle the origin load. What optimization is required? | fail→pass | 9,648 | 10,276 | +7% | 1 | 1 | 0% | 1,652 | 2,027 | +23% | 0 | 0 | — |
▸case-13 User checkout API takes 6.5 seconds to return a response because it generates a PDF invoice, sends a transactional email via SMTP, and calls a third-party CRM webhook synchronously inside the request handler. How should this endpoint be optimized? | pass→pass | 13,672 | 14,775 | +8% | 1 | 1 | 0% | 2,318 | 2,812 | +21% | 0 | 0 | — |
▸case-14 Searching users by email using `WHERE email LIKE '%@gmail.com'` on a 5M row table takes 8 seconds despite having a standard B-tree index on `email`. A developer suggests adding a read replica. What is the correct optimization? | pass→pass | 12,435 | 11,381 | -8% | 1 | 1 | 0% | 2,127 | 2,193 | +3% | 0 | 0 | — |
▸case-15 An internal REST API returns a 12MB uncompressed JSON payload for dashboard reporting, consuming high network bandwidth over cellular connections. Developers want to paginate the endpoint into 1,000 separate calls. What immediate optimization delivers quick impact? | pass→pass | 8,664 | 10,956 | +26% | 1 | 1 | 0% | 1,429 | 2,048 | +43% | 0 | 0 | — |
▸case-16 Our Kubernetes cluster running 50 microservices costs $18,000/month. Metrics show node CPU utilization averages 8% and memory 22% because pod resource requests are set to 8 vCPU and 32GB RAM per pod. What is the recommended cost optimization? | pass→pass | 13,331 | 13,917 | +4% | 1 | 1 | 0% | 2,343 | 2,568 | +10% | 0 | 0 | — |
▸case-17 Node.js application memory usage steadily climbs over 24 hours until the process crashes with Out Of Memory (OOM). Profiling shows unclosed database client instances inside an unhandled error path. How should this be fixed? | fail→fail | 12,158 | 15,734 | +29% | 1 | 1 | 0% | 2,162 | 2,639 | +22% | 0 | 0 | — |
▸case-18 A React web application JavaScript bundle size is 8.5MB uncompressed, resulting in an 11-second Time to Interactive on mobile networks. The team imports entire libraries like `lodash` and `date-fns` for single helper functions. What is the optimization strategy? | pass→pass | 13,378 | 14,652 | +10% | 1 | 1 | 0% | 2,735 | 3,092 | +13% | 0 | 0 | — |
▸case-19 A heavily updated PostgreSQL table of user sessions experiences severe query slowdown over weeks despite having proper B-tree indexes. Disk usage for the table is 40GB even though row count remains constant at 100,000 rows. What action fixes this? | pass→pass | 7,967 | 8,893 | +12% | 1 | 1 | 0% | 1,544 | 1,800 | +17% | 0 | 0 | — |
▸case-20 A web client loads 80 individual small JSON asset files via HTTP/1.1, hitting browser per-domain connection limits (6 parallel connections) and creating severe head-of-line blocking. How should the backend network layer be optimized? | pass→pass | 13,981 | 12,513 | -10% | 1 | 1 | 0% | 2,400 | 2,430 | +1% | 0 | 0 | — |
▸case-21 We have three performance issues: (A) Missing DB index on user search causing 5-second queries for all users (1 line change), (B) Uncompressed 50MB static video on landing page (30 min fix), (C) Complete rewrite of auth service from Node.js to Rust to save 5ms latency (3 months effort). Rank these in an optimization backlog. | pass→pass | 10,155 | 8,222 | -19% | 1 | 1 | 0% | 1,634 | 1,799 | +10% | 0 | 0 | — |
▸case-22 We just applied composite database indexes and enabled Redis response caching on our checkout microservice. What concrete steps should we present to validate that the performance optimization was successful? | pass→pass | 16,393 | 12,916 | -21% | 1 | 1 | 0% | 2,579 | 2,415 | -6% | 0 | 0 | — |