zhoujiexiong commented on PR #11415: URL: https://github.com/apache/apisix/pull/11415#issuecomment-2264944953
> > this bug only occurs when both `healthy` and `max_breaker_sec` are configured together? > > No,This bug is easy to reproduce.It can be reproduced with the default configuration.But it needs to be explained in conjunction with the api-breaker plug-in code. As shown in the following picture:  In an unhealthy state,The request is then retried in 2, 4, 8, 16 ... seconds until the max_breaker_sec.In this process, there is often a situation:when the specified unhealthy state has not been reached, the breaker is triggered. The reason is that when the specified number of unhealthy states is not reached, lasttime + breaker_time will occasionally appear to be greater than or equal to ngx.time().This causes a breaker to occur. The solution is shown in the following figure:  Suppose there are 4 requests inbound, 3 failed, breaker triggered, then the 4th increase the `unhealthy_count` at `log_phase`, the modification of the PR would exit the breaking state? -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
