Hi all,
I would like to inform you that we have slightly changed our thoughts about
the implementation
of the Token Bucket algorithm. Our initial idea was to change our existing
Rate to behave like
a Token Bucket. That works as we expected but we have realized that the
value of the Rate is
not
Hi all,
Just a quick update. We have made good progress regarding the implementation
of this KIP. The major parts are already in trunk modulo the new "rate"
implementation
which is still under development.
I would like to change the type of the `controller_mutations_rate` from a
Long to
a
Hi all,
The vote has passed with 5 binding votes (Gwen, Rajini, Mickael, Jun and
Colin)
and 2 non-binding votes (Tom, Anna).
Thank you all for the fruitful discussion! I'd like to particularly thank
Anna who has
heavily contributed to the design of this KIP.
Regards,
David
On Fri, Jun 12, 2020
+1. Thanks, David!
best,
Colin
On Thu, Jun 11, 2020, at 23:51, David Jacot wrote:
> Colin, Jun,
>
> Do the proposed error code and the updated KIP look good to you guys? I’d
> like to wrap up and close the vote.
>
> Thanks,
> David
>
> Le mer. 10 juin 2020 à 14:50, David Jacot a écrit :
>
Hi, David,
Thanks for making those changes. They look fine to me. +1
Jun
On Thu, Jun 11, 2020 at 11:51 PM David Jacot wrote:
> Colin, Jun,
>
> Do the proposed error code and the updated KIP look good to you guys? I’d
> like to wrap up and close the vote.
>
> Thanks,
> David
>
> Le mer. 10
Colin, Jun,
Do the proposed error code and the updated KIP look good to you guys? I’d
like to wrap up and close the vote.
Thanks,
David
Le mer. 10 juin 2020 à 14:50, David Jacot a écrit :
> Hi Colin and Jun,
>
> I have no problem if we have to rewrite part of it when the new controller
>
Hi Colin and Jun,
I have no problem if we have to rewrite part of it when the new controller
comes
out. I will be more than happy to help out.
Regarding KIP-590, I think that we can cope with a principal as a string
when the
time comes. The user entity name is defined with a string already.
Hi, Colin,
Good point. Maybe sth like THROTTLING_QUOTA_VIOLATED will make this clear.
Hi, David,
We added a new quota name in the KIP. You chose not to bump up the version
of DESCRIBE_CLIENT_QUOTAS and ALTER_CLIENT_QUOTAS, which seems ok since the
quota name is represented as a string. However,
On Tue, Jun 9, 2020, at 05:06, David Jacot wrote:
> Hi Colin,
>
> Thank you for your feedback.
>
> Jun has summarized the situation pretty well. Thanks Jun! I would like to
> complement it with the following points:
>
> 1. Indeed, when the quota is exceeded, the broker will reject the topic
>
Hi, David,
Sounds good then.
Thanks,
Jun
On Tue, Jun 9, 2020 at 10:59 AM David Jacot wrote:
> Hi Jun,
>
> Both are already in the KIP, see "New Broker Configurations" chapter. I
> think
> that we need them in order to be able to define different burst for the new
> quota.
>
> Best,
> David
>
Hi Jun,
Both are already in the KIP, see "New Broker Configurations" chapter. I
think
that we need them in order to be able to define different burst for the new
quota.
Best,
David
On Tue, Jun 9, 2020 at 7:48 PM Jun Rao wrote:
> Hi, David,
>
> Another thing. Should we add
Hi, David,
Another thing. Should we add controller.quota.window.size.seconds and
controller.quota.window.num
or just reuse the existing quota.window.size.seconds and quota.window.num
that are used for other types of quotas?
Thanks,
Jun
On Tue, Jun 9, 2020 at 10:30 AM Jun Rao wrote:
> Hi,
Hi, David,
Thanks for the KIP. The name QUOTA_VIOLATED sounds reasonable to me. +1 on
the KIP.
Jun
On Tue, Jun 9, 2020 at 5:07 AM David Jacot wrote:
> Hi Colin,
>
> Thank you for your feedback.
>
> Jun has summarized the situation pretty well. Thanks Jun! I would like to
> complement it with
Hi Colin,
Thank you for your feedback.
Jun has summarized the situation pretty well. Thanks Jun! I would like to
complement it with the following points:
1. Indeed, when the quota is exceeded, the broker will reject the topic
creations, partition creations and topics deletions that are
On Mon, Jun 8, 2020, at 14:41, Jun Rao wrote:
> Hi, Colin,
>
> Thanks for the comment. You brought up several points.
>
> 1. Should we set up a per user quota? To me, it does seem we need some sort
> of a quota. When the controller runs out of resources, ideally, we only
> want to penalize the
Hi, Colin,
Thanks for the comment. You brought up several points.
1. Should we set up a per user quota? To me, it does seem we need some sort
of a quota. When the controller runs out of resources, ideally, we only
want to penalize the bad behaving applications, instead of every
application. To
Hi David,
Thanks for the KIP.
I thought about this for a while and I actually think this approach is not
quite right. The problem that I see here is that using an explicitly set quota
here requires careful tuning by the cluster operator. Even worse, this tuning
might be invalidated by
+1 (not binding)
Thanks for the KIP!
-Anna
On Thu, Jun 4, 2020 at 8:26 AM Mickael Maison
wrote:
> +1 (binding)
> Thanks David for looking into this important issue
>
> On Thu, Jun 4, 2020 at 3:59 PM Tom Bentley wrote:
> >
> > +1 (non binding).
> >
> > Thanks!
> >
> > On Wed, Jun 3, 2020 at
+1 (binding)
Thanks David for looking into this important issue
On Thu, Jun 4, 2020 at 3:59 PM Tom Bentley wrote:
>
> +1 (non binding).
>
> Thanks!
>
> On Wed, Jun 3, 2020 at 3:51 PM Rajini Sivaram
> wrote:
>
> > +1 (binding)
> >
> > Thanks for the KIP, David!
> >
> > Regards,
> >
> > Rajini
>
+1 (non binding).
Thanks!
On Wed, Jun 3, 2020 at 3:51 PM Rajini Sivaram
wrote:
> +1 (binding)
>
> Thanks for the KIP, David!
>
> Regards,
>
> Rajini
>
>
> On Sun, May 31, 2020 at 3:29 AM Gwen Shapira wrote:
>
> > +1 (binding)
> >
> > Looks great. Thank you for the in-depth design and
+1 (binding)
Thanks for the KIP, David!
Regards,
Rajini
On Sun, May 31, 2020 at 3:29 AM Gwen Shapira wrote:
> +1 (binding)
>
> Looks great. Thank you for the in-depth design and discussion.
>
> On Fri, May 29, 2020 at 7:58 AM David Jacot wrote:
>
> > Hi folks,
> >
> > I'd like to start the
+1 (binding)
Looks great. Thank you for the in-depth design and discussion.
On Fri, May 29, 2020 at 7:58 AM David Jacot wrote:
> Hi folks,
>
> I'd like to start the vote for KIP-599 which proposes a new quota to
> throttle create topic, create partition, and delete topics operations to
>
22 matches
Mail list logo