Learn how service performance optimization tasks are decomposed and managed in this 45-minute conference talk delivered in Russian at DevOops. Explore comprehensive aspects of performance engineering including testing methodologies, profiling techniques, and benchmarking practices. Discover approaches for designing service metrics, establishing SLO connections, analyzing failures, and implementing performance optimization workflows. Gain practical insights into automating performance monitoring within CI/CD pipelines while understanding when and how performance considerations become critical for service reliability engineering.