alibaba / alibaba/Sentinel

会在某一个1s突然出现过多失败次数,造成错误熔断

Open
#2,873 1 comment 0 reactions 0 assignees View on GitHub
area/circuit-breaking
Dominant language
Java
Stars
23.1k
Forks
8.1k
PR merge metrics
No merged PRs in 30d

Description

## Issue Description

Type: 利用dubbo rpc框架,采用的随机权重负载均衡策略,其中某个下游服务器有8个节点,突然有个节点频繁old gc出现响应超时。consumer单台机器 qps 13左右,设置的降级规则,为失败率50%进行熔断,按说1/8的节点,不会到达50的失败率。但是看sentinel metric日志,会在某一个1s窗口突然出现10次左右失败统计,造成失败率超过50%进行熔断,所以反而把服务质量变得更差了,求解

![image](https://user-images.githubusercontent.com/7530863/189374712-57e2655e-4d31-480b-a098-fdb45ab958c0.png)

Contributor guide

Open the contributing guide

Research direction

Start with Sentinel's metric logging and circuit-breaker failure-rate calculation; the issue names no source files or tests. Reproduce the one-second window using roughly 13 QPS, eight downstream nodes, one node timing out during old GC, and a 50% failure threshold. Done means explaining or correcting the unexpected burst of about 10 failures and its resulting circuit break.

Written by the indexing model from the issue text.

Assessment

Tech stack
java
Domain
backend, distributed-systems, observability
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.