For some high performance environments, I would like to see microsecond or
nanosecond precision on metrics whenever possible. Even better would be
some sort of histogram of individual events so we could see the
variability.
On 12/21/15, 9:27 PM, "Alexey Pirogov" wrote:
>Ismael, thanks for reply
Ismael, thanks for reply.
Jire created https://issues.apache.org/jira/browse/KAFKA-3028.
Thank you,
Alexey
Hi Alexey,
Could you please report a bug in JIRA for the NaN result? We should handle
that better.
Thanks,
Ismael
On Mon, Dec 21, 2015 at 9:12 AM, Alexey Pirogov
wrote:
> I'm looking for help with a question regarding measuring of producer
> request latency.
> I expected that "producer.request
I'm looking for help with a question regarding measuring of producer
request latency.
I expected that "producer.request-latency-avg/max" will do a good job for
me. But seems that if latencies less than 1ms in most cases, this metric
will emit NaN(as it doesn't support float values).
We need this me
