Regarding the issue of tracing performance bottlenecks in concurrent request chains.

zeromicro / go-zero

A cloud-native Go microservices framework with cli tool for productivity.

https://go-zero.dev

MIT License

29.37k stars 3.97k forks source link

Regarding the issue of tracing performance bottlenecks in concurrent request chains. #4350

Open JackGod001 opened 2 months ago

JackGod001 commented 2 months ago

设备 Apple M1 Max docker 中分配了9核cpu jaeger 中一个简单的api /v1/user-center/user 有的时候有redis部分,有的时候没有为什么? grafana 中面板参数如下查看到的状态码 {{code}} - 请求速率 (每秒) 请求量只有2500左右 sum by (code) (rate(http_server_requests_code_total[$rate_interval])) redis 的面板设置 rate(redis_client_requests_duration_ms_count[5m]) 最高800 不知道问题出在哪里了

JackGod001 commented 2 months ago

The Apple M1 Max device in docker is allocated with 9 CPU cores. In Jaeger, there is a simple API /v1/user-center/user. Sometimes there is a Redis component, and sometimes there isn't. Why is that? In Grafana, the panel parameters show that the request rate (per second) is only around 2500 requests. sum by (code) (rate(http_server_requests_code_total[$rate_interval])) Redis panel settings rate(redis_client_requests_duration_ms_count[5m]) Up to 800. I don't know where the problem is.