Fix P2P + P2S statuc route behaviour after additions on Azure VPN Gateways
At present, in a situation where a single Azure VPN Gateway is used as both a P2P and P2S device, adding fixed downstream routes to the Local Gateway and P2S endpoint (both split and forced tunnel modes) results in VPN endpoints being unable to reliably route to the downstream locations. BGP is not in use.
At present, a reset of the VPNGW is required when adding new routes in order for them to work correctly (even though routing works correctly from any servers inside an attached VNET, and published routes propagate correctly to the P2S clients via intune or manual route table changes on the device.).
This appears to be a bug in the VPNGW, as it is reproducible and I can find no reference to having to reset the appliance to make changes function correctly in any Microsoft Documentation.
MS support failed to diagnose this issue when a ticket was raised, and I finally worked it out myself after much investigation, trial and error.