Create an API that helps understand the latency of chef infra server with its external connections like elasticsearch and postgres.
In recent support calls, it was found very difficult to diagnose if the slowness observed was due to chef-server or elastic-search or the latency between the two.
Having some instrumentation around monitoring the system to detect slowness would be very helpful in these scenarios.
It could be implemented with some counters (in prometheus if possible)
1) the number of requests to es
2) time taken per request
3) request success and failure counts