Hi I am trying to run parallel processing with crew ai so that it runs faster but when running simultaneously it looks like the outputs and the inputs get all mixed up.
Do you have experience running ThreadPoolExecutor successfully with crewai and if so how did you do it?
I have considered using kickoff_for_each but I also saw that the tracking of outputs is just as difficult, especially for hundreds of different inputs.
Thank you!
Rachel
Try using the crewai hosting it is setup to scale. It might solve you issue till you can build out your own environment