When there is just a single level, even with lots (+17,000 files), the watch node works surprisingly well but as soon as I add subdirectories and turn on "Watch sub-directories recursively", deploying then takes minutes for any change in the flows. Any alternative nodes or methodology would be appreciated. The files are on a network drive if that makes any difference.
I tried just creating sub folders (21 levels) and was quick. Did a check and one of the locations has 30,000 files and 35,000 folders. Annoyingly I'm only interested in a small subset of files that have a naming convention in the folder. May try getting my hands dirty and creating a version with inclusion rules but not sure if that will make it any better, maybe the sheer numbers is the issue.
Sorry I thought this was a run time issue not a deploy issue. What option for deploying are you using? Is it Full? If so, try using Modified Flows instead