Thanks for responding to feedback. Had a few more points.
> The idle_time can be part of active_time in specific scenarios - > particularly when the controller has sent records to the Raft layer and is > awaiting commitment confirmation, but not when simply waiting for new > events to enter the queue. I'm a bit confused about why "idle-time" is not a time-complement of "active-time"? IE this is the time spent "not processing any events at all?" Would it be useful to clarify what was occurring while waiting? IE "raft layer" or "no events in the queue" by adding a dimension to the metric? Type=Gauge The metric type is specified as a Gauge. IUUC gauges are usually for taking instantaneous measurements - for example "heat of oven = 80C" or "Memory Usage = 500Mb". These are measurements which do not depend on things occurring in the past. It seems this metric depends on things which previously happened (past event-processing) and so a gauge may not be suitable. EventQueueTimeMs <https://kafka.apache.org/documentation/#kraft_controller_monitoring>, for example, is a histogram. In a similar vein, has tracking "idleness" as discrete periods of time (sort of like pseudo-events where nothing happens) + histogram rather than a ratio of "idle / (active + idle)" been considered? Best, Jonah On Thu, Jul 10, 2025 at 3:09 PM Mahsa Seifikar <mseifi...@confluent.io.invalid> wrote: > Hi Jonah and Kevin, > > Thanks for your comments. I have now updated the KIP to address your > feedback. > > Please let me know if you have any further questions. > > Best, > Mahsa Seifikar > > On Thu, Jul 3, 2025 at 4:40 PM Mahsa Seifikar <mseifi...@confluent.io> > wrote: > > > Hello all, > > > > I wrote a short KIP to add a new metric for controller thread idleness. > > > > Here is the link: > > > > > https://cwiki.apache.org/confluence/display/KAFKA/KIP-1190%3A+Add+a+metric+for+controller+thread+idleness > > > > Thanks, > > Mahsa Seifikar > > > > > > >
