Please add an option to the Erwin Harvester node to keep Govern models even if the XML file is not in the data directory at runtime.
Currently, we must keep XML files of all models in the data directory or they will be removed from Govern. I found that the Harvester will process every model each time it is run and also compare to Govern. We have several models that are very large and processing all of them takes a lot of compute resources. I've even had Analyze crash because the server ran out of memory.
While there may be some cases where removing models from Govern via the Harvester would be beneficial, for us, it is usually detrimental.