Flyte enables production-grade orchestration for machine learning workflows and data processing created to accelerate local workflows to production.

Flyte

Is there a way to access `max_parallelism` from within a task to make a batching decision at the point of a fan-out? Or is this difficult to figure out because I shouldn't be doing it? :sweat_smile:

I'm getting the impression that I am trying to do something bad since the FlyteRemote docs also <https://www.union.ai/docs/flyte/user-guide/development-cycle/flyte-remote/|discourage use in Tasks>
&gt; _Do not use `FlyteRemote` within task code._

the only other similar documented approach is passing a `chunk` argument around? <https://www.union.ai/docs/flyte/user-guide/programming/nested-parallelism/> -- but seems like you could use `max_parallelism` to determine your chunk size more appropriately.

Yes using control plane in tasks is discouraged as it may not be as performant 

<@U08C7U122UV> we are working on something new that solves this problem at the core - would you be interested in talking and sharing thoughts 