option to restrict call cache size

[MatthewMah]

I run cromwell on a SLURM cluster with a postgres database for call caching. Call caching mostly works fine, but there is no way to constrain the size of the cache. Eventually the available space (200GB) for the postgres database fills up, and all cromwell instances using the call cache fail. My workaround is to periodically stop all workflows and truncate the call caching database.

I imagine implementing an option to restrict the number of cache entries should be straightforward for a Guava-based cache; if someone could help roadmap, I may be able to submit a pull request in the future.

For my purposes, I assume that all cromwell instances agree on the size.