Using pprof to diagnose performance issues in Golang services

Jan 8, 2025##golang #pprof #web59

AI Translation

This post is translated from Chinese into English through AI.View Original

AI-generated summary

If you notice high CPU or memory usage or slow response times in your service, pprof can help identify problematic code. To use pprof, simply import the library and add a few lines to your server code. Start a goroutine to listen on a specific port for profiling data. After collecting data for 30 seconds, you can analyze it using the command `go tool pprof -http :8081 ./profile` and access the results at `http://localhost:8081/ui/`. You can also select a flame graph for visualization.

If you find that the CPU/memory usage is too high, or the request response is slow, and you want to know which part of the code is causing the problem, pprof is used to help you with this.

First, you need to import the pprof library in advance. It's simple; just add a few lines of code to your server code.

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

func main() {
	go func() {
		http.ListenAndServe("localhost:8080", nil)
	}()
	// Main process
}

Click on the profile in the image below, which will collect 30 seconds of data and save it to the profile file.

Pasted image 20250108223541

Then run

go tool pprof -http :8081 ./profile

Open http://localhost:8081/ui/

Pasted image 20250107105500

In the upper left corner, you can select the legendary flame graph (Flame Graph):

Pasted image 20250107105539