Hi Yann,

Yes, it is possible to add such a feature. It will be less efficient as
it will require to manage the container buffer dynamically (to grow and
shrink it, define an upper boundary for growth, etc.) and require a bit
of coding. I can add to my todo list for 1.7.1 (as 1.7.0 code will go
freeze this week and be released soon). Can you set a priority for this?
Shall i say in the next month, sometime in October, how does it sound?

Paolo

On Tue, Sep 19, 2017 at 07:42:36AM +0000, Yann Belin wrote:
> Thanks Paolo,
> 
> >From your opinion, would such feature be difficult (possible) to implement?
> I didn't look into the code yet, I didn't program in C since high school
> and my skills are quite rusty.
> 
> Cheers,
> 
> Yann
> 
> On Fri, Sep 15, 2017 at 6:38 PM Paolo Lucente <pa...@pmacct.net> wrote:
> 
> >
> > Hi Yann,
> >
> > I confirm you can't do that with AMQP as the only knob pmacct gives
> > you is the size-based amqp_multi_values. Although not part of your
> > question, with Kafka and you may choose not to leverage the pmacct
> > knob, kafka_multi_values, and use instead batch.num.messages (ie.
> > amount of messages you want to batch before sending to the broker) or
> > queue.buffering.max.ms (queue for some given amount of time, if for
> > example latency is the constraint) offered by librdkafka.
> >
> > Paolo
> >
> > On Thu, Sep 14, 2017 at 04:12:45PM +0200, Yann Belin wrote:
> > > Hello,
> > >
> > > Does anyone know if there is a way to control the maximum number of
> > > rows sent in an AMQP message?
> > >
> > > amqp_multi_values allows me to do that in an approximate way
> > > (row_size/message_size) but I need to have a finer control on that,
> > > which is not dependent of variations of row size.
> > >
> > > Thanks in advance,
> > >
> > > Yann
> > >
> > > _______________________________________________
> > > pmacct-discussion mailing list
> > > http://www.pmacct.net/#mailinglists
> >
> > _______________________________________________
> > pmacct-discussion mailing list
> > http://www.pmacct.net/#mailinglists
> >

_______________________________________________
pmacct-discussion mailing list
http://www.pmacct.net/#mailinglists

Reply via email to