Release 7.2.0

Improved performance of stream recovery on edge with rapid-republishing over cluster
Removed GC run loop which could cause stream irregularities at 30 second intervals
Master HLS record can be created for ABR multi-variant recordings
Fixed - issue where server did not run correctly on newer Windows Server versions
Autoscale - Fixed - Problem trying to publish to a sub-scope via Stream Manager Proxy
Autoscale - Fixed - Missing streams on all of the edges with fast transcoder republish
Autoscale - Updated logic to prevent child nodes from using temporary ids of parent nodes which are being replaced
Autoscale - Fixed - Exception thrown on region subscribe request if the stream doesn't exist on an edge in the specified region