Best Practices for Running 10k+ Rows in Clay Batches

Looking for the best suggestion on how to run a Clay table of 10k+ in 50 row batches? The API access to a platform we are using via Clay is also being used by the product itself at our company, So I dont want to overload the system and want to just do it in small batches vs. overloading it Any suggestions? Doing it 1x1 or in groups of 50 is going to take forever. So just curious on best tips for this