In the current implementation of parallel checkpoint we internally loop in the function __checkpoint_parallel_thread_run, there is likely a small performance benefit to this but whether it is necessary or not has not yet been validated.

In this ticket we should remove the internal loop and check whether performance gets worse as a result. To do so run the relevant evergreen tasks:

evergreen patch -p sys-perf --variants disagg-perf-10-node.arm.aws --tasks ycsb.load_0128thread.2024-05

Checking whether ops_per_sec gets worse or better, and also checking if the number of pages reconciled per second by parallel checkpoint improves or not. Given this optimization is relatively clear, and safe this ticket is low priority.