Github user kl0u commented on the issue:
https://github.com/apache/flink/pull/4654
Hi @zjureel ! Thanks for the work!
This can be a temporary fix, but I was thinking more of a long term one
where there is no limit.
The problems that I can find with such temporary solution are:
1) if we add this as a configuration parameter, and then remove it when we
have a proper fix, this may result in confusion for the users.
2) with this fix, the aggregator is going to allocate the specified limit
every time there is an element to parse, even though the element may be small.
This can be a waste of resources.
What do you think @zentol ?
---
