Github user karanmehta93 commented on the issue:
https://github.com/apache/zookeeper/pull/307
> Can we at least make it so that the process of "push a metric to a
buffer, have a thread that wakes up periodically and flushes information out of
that buffer" is usable by multiple parts of the system, instead of coupling it
to the one metric of request time?
I am ready to put in the effort of making this framework more generic, if
you feel its worth the time and effort to put into it. I would also like to
hear others opinions on this.
The inspiration for this JIRA came from HBase, where it has been used to
log slow requests at the basic level. I feel that it can be useful to keep
track of latency spikes that are seen in production. @apurtell can suggest
use-cases.
---
