How did you measure the reduction of CPU load by 35% and latency?

Reducing CPU load and latency is usually measured through profiling and monitoring:

• For CPU, use profilers such as pprof in Go. It allows collecting data on which functions consume CPU and for how long.
• For latency, response time metrics are collected via monitoring tools (Prometheus, Grafana).

Example of using pprof for CPU profiling:

import (
    "net/http"
    _ "net/http/pprof"
)

func main() {
    go func() {
        log.Println(http.ListenAndServe("localhost:6060", nil))
    }()
    // main application logic
}

After starting, you can connect to http://localhost:6060/debug/pprof/profile to get the CPU profile.

By comparing profiles and metrics before and after optimization, you can assess the reduction in CPU load and latency.