QA will run load tests against Compute Engine services that use Cloud Bigtable. Which three requirements should the load-testing plan include?

Validating Cloud Bigtable performance is crucial because it's a key component under test, and its scalability directly impacts the application's overall performance. Using a separate Google Cloud project for load testing isolates the testing environment from production, preventing accidental impact on live services and ensuring resource quotas aren't shared, which could skew results. Instrumenting the load-testing tool and target services with detailed logging and metrics provides essential data for analyzing performance bottlenecks and understanding system behavior under load. Running load tests regularly against production is risky and can disrupt live users. Ensuring all third-party dependencies can tolerate high load is a good practice but not a direct requirement for the load-testing plan itself, which focuses on the application under test. Instrumenting production to record every transaction for replay is generally impractical and resource-intensive for load testing.