You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@kafka.apache.org by Arunkumar <pm...@yahoo.com.INVALID> on 2018/06/19 19:03:32 UTC
ISR shrink rate & IsrExpandsPerSec issue
Hi All,
I am seeing IsrShrinksPerSec & IsrExpandsPerSec increases (>0) when one or more brokers goes down and comes back up into the cluster. But the values should get back to 0 once the servers catches up, is my understanding after reading most of the documents. But in our production environment (3 node cluster) these values does not reduce and keeps increasing. This keeps throwing alerts every few seconds. Any insight is highly appreciated.
Thanks in advanceArunkumar Pichaimuthu, PMP
Re: ISR shrink rate & IsrExpandsPerSec issue
Posted by Arunkumar <pm...@yahoo.com.INVALID>.
Make sense, Thank you Mani. I appreciate your time and effort.
ThanksArunkumar Pichaimuthu, PMP
On Wednesday, June 20, 2018, 1:18:24 AM CDT, Manikumar <ma...@gmail.com> wrote:
These metrics are meter type metrics, which tracks count, mean rate, 1-,
5-, and 15-minute moving averages.
You maybe observing count measure, which gives number of events which have
been marked.
You can try monitoring 1/5/15 miniture averages.
On Wed, Jun 20, 2018 at 12:33 AM Arunkumar <pm...@yahoo.com.invalid>
wrote:
> Hi All,
> I am seeing IsrShrinksPerSec & IsrExpandsPerSec increases (>0) when one
> or more brokers goes down and comes back up into the cluster. But the
> values should get back to 0 once the servers catches up, is my
> understanding after reading most of the documents. But in our production
> environment (3 node cluster) these values does not reduce and keeps
> increasing. This keeps throwing alerts every few seconds. Any insight is
> highly appreciated.
>
> Thanks in advanceArunkumar Pichaimuthu, PMP
Re: ISR shrink rate & IsrExpandsPerSec issue
Posted by Manikumar <ma...@gmail.com>.
These metrics are meter type metrics, which tracks count, mean rate, 1-,
5-, and 15-minute moving averages.
You maybe observing count measure, which gives number of events which have
been marked.
You can try monitoring 1/5/15 miniture averages.
On Wed, Jun 20, 2018 at 12:33 AM Arunkumar <pm...@yahoo.com.invalid>
wrote:
> Hi All,
> I am seeing IsrShrinksPerSec & IsrExpandsPerSec increases (>0) when one
> or more brokers goes down and comes back up into the cluster. But the
> values should get back to 0 once the servers catches up, is my
> understanding after reading most of the documents. But in our production
> environment (3 node cluster) these values does not reduce and keeps
> increasing. This keeps throwing alerts every few seconds. Any insight is
> highly appreciated.
>
> Thanks in advanceArunkumar Pichaimuthu, PMP