http health check unstable behavior

[maxlavr]

Title: why http health check status 200, but health_flags failed_active_hc

Description:

Repro steps:

envoy v1.20.7

Admin and Stats Output:

/clustes cluster's upstreams constantly changes from healthy to unhealthy

Config:

Logs:

screenshot in description

Either I don’t fully understand the logic of processing health checks or there is another explanation why the response returned 200, but flag = failed_active_hc, help me please

[adisuissa]

If I read the code correctly the debug log line prints the host state before processing the result.