Sounds good.
Just pointing out that the metrics scope likely overlap with
https://issues.apache.org/jira/browse/KAFKA-3556, so may better document
that in the wiki page.
Guozhang
On Mon, Jul 29, 2019 at 2:48 PM Jason Gustafson wrote:
> Hi Guozhang,
>
> I have added metrics to the KIP. Please
Hi Guozhang,
I have added metrics to the KIP. Please take a look. This gave me an excuse
to also add a metric for the group rebalance rate, which probably would
have made detecting KAFKA-8653 easier.
Since this is a relatively straightforward KIP, I will go ahead and start a
vote later this week
Thanks for the replies Jason!
2. No I do not see any problems, just trying to understand how restrict we
are applying this rule :) Piggy-backing on the existing background thread
and check interval mechanism means we are not "eagerly" expiring either,
but I think this is fine.
Guozhang
On
1. Fixed, thanks!
2. Yes, that is what I was thinking. Do you see any problems?
3. Good point. Do you think a meter for expired and deleted offsets would
be sufficient?
4. I considered it. I thought that might be a little dangerous for dynamic
groups which have subscriptions changing. If the
Hi Jason,
Thanks for the KIP! I've made a pass on it and here are few comments:
1. " before the clients which ." --> incomplete sentence?
2. " Any committed offset for a partition which is not currently subscribed
to is subject to expiration." --> this may be an implementation detail, but
are
Hi All,
I have a short KIP to add an api for consumer offset deletion:
https://cwiki.apache.org/confluence/display/KAFKA/KIP-496%3A+Administrative+API+to+delete+consumer+offsets.
Please take a look and let me know what you think.
Thanks,
Jason