We need to bring up the code coverage within the RTS module to be as high as possible to ensure that we are testing the entire code paths. We should cover as much of the module as possible, perhaps only with the exception of panic states.

For example, WT-13352 identified that the workload generator was not producing checkpoints with evicted prepared data, which, if we had ready, could have identified the issue in WT-13252 before it came as a hot BF.

This work would likely need to be split into several tickets due to its large scope.