Circuit breaker not triggered with header on endpoint configuration

[exalate-issue-sync[bot]]

Describe the bug :

The client was testing whether Healtcheck works properly in APIM 3.20.15. They encountered an issue where they were expecting a 503 error, but they were either getting 200 or 404 depending on whatever endpoint they provided. Their purpose was to make healthcheck delete an endpoint after it wasn't working properly. But after a call with them, we found out that if they enable system proxy on endpoint level, that is when the error occurs, so the headers don't matter.

This was later escalated to our developers to which they responded with this: When adding a health check with a 30s rate, the endpoint status is not updated as expected to DOWN if the health check assertion fails. Hence, health checks are shown as failed in the health check dashboard, but the endpoint continues to handle requests addressed to it. Lowering the rate to 5s restore the expected behaviour. (503 status if HC fails, 200 if HC succeeds)

Maybe a sync problem between the scheduler rate and the health check rate.

_{* Please see Zendesk Support tab for further comments and attachments.}

[exalate-issue-sync[bot]]

This issue has been fixed, a new version will be available soon