Allow increase of `maxiter` for OptimizationDatasets

[dotsdl]

Is your feature request related to a problem? Please describe.

It is often the case that the initial maxiter setting for geometry optimization datasets is set too low on submission, resulting in many OptimizationRecords that will never complete unless the maxiter is set higher.

Describe the solution you'd like

Because this setting is entirely a compute parameter, and has no bearing on the final result (except for completion, as noted), we would like to be able to alter this on an existing OptimizationDataset, or at the very least increase it, after submission.

This may be tricky, as existing tasks associated with each OptimizationRecord would need to be either altered or replaced with the new maxiter.

Describe alternatives you've considered

Currently, we often re-submit the compute spec with a higher number of iterations specified on the dataset. This is rather wasteful, as it will re-compute for cases already completed in addition to those that haven't.

Additional context

[bennybp]

I don't see a way to do this with individual records. However, you can now set additional_keywords on entries in an OptimizationDataset. These are merged with the specification keywords when creating records for a particular entry. So those records would be recalculated, but at least already-completed entries wouldn't be recomputed.