/health endpoint about openwhisk-package-kafka HOT 4 CLOSED

Other statistical information is also valuable (e.g. average time to process incoming messages, size of messages processed, etc.) but these might be better handled by statistical analysis of the log messages, rather than actively computing these inside the feed service provider itself.

from openwhisk-package-kafka.

I suspect this will need to be implemented as a separate thread that wakes up every now and then, computes and records the required information.

from openwhisk-package-kafka.

@bjustin-ibm, how do you feel about having a separate thread running that periodically outputs the health information to the logs?

from openwhisk-package-kafka.

@dubeejw I suspect that will be necessary in order to provide "average" stats. However, it is possible that the /health endpoint could simply return the current values, and some other monitoring system will handle computing averages by periodically polling the /health endpoint and recording the results.

For now, let's just make it return the instantaneous values. Later, if we have a need, we can compute/record values using a separate thread.

from openwhisk-package-kafka.