markusthoemmes commented on issue #2482: Add container pool health status in invoker URL: https://github.com/apache/incubator-openwhisk/pull/2482#issuecomment-315290892 Let me explain: The health ping is solely to check if the invoker is there. That one has nothing to do with any state inside the invoker as you pointed out. We do have a second mechanism named active ack. It's used to fastly send back the response, which is then sent to the user. The controller checks those active ack messages for system errors (not application errors). If for a given timeframe one invoker sends too many subsequent system errors the invoker is considered unhealthy and checked again. If missing active acks are not caught by that, as @rabbah pointed out, we have a bug there. And important bit to note is: The piece of code you instrumented does not necessarily indicate unhealthyness. That code can easily be reached even in a fully healthy scenario. ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected]
With regards, Apache Git Services
