Craig,
If you have a timestamp set as an attribute on the processor, then this is kind
of possible.
Have a regular MergeContent processor, with "Maximum Group Size" set to 1 mb,
set "Max Bin Age" to 3 min; you may need to tweak settings to get the right
cadence, but these are generally the set
...from the description it isn't clear what you're trying to achieve so
lets first try to expand the detail on the use case.
We should distinguish whether you're wanting to 'combine various objects in
a data stream together on some time bound' from 'processing various objects
in a data stream to m
Joe
Thanks for the options.
Kafka Flink option is interesting but to big at the moment.
I’ll build a custom processor for this occasion.
Much appreciated!
Craig Knell
> On 4 Jun 2019, at 21:56, Joe Witt wrote:
>
> ...from the description it isn't clear what you're trying to achieve so
Hi Craig,
I was doing something similar couple of years ago. Built custom
processor and found out, that the problem is much bigger then I
thought.
In custom processor, the onTrigger is called from different threads,
in parallel. So you have to maintain synchronized or concurrent data
structures in