Online Managed Instance Scale operations
Scaling a managed instance takes the instance completely offline which means scaling can only be down during outage windows which defeats the purpose of a PaaS solution. Even scaling storage capacity which is a billing scale up rather than a physical one seems to take the instance down for up to 3 minutes. Then the instance has to start from scratch with query plans and cache.
Is there no way to have the new instance primed and cut over to rather than killing and starting up from scratch?
Experienced peaks of 18,5h to scale up or down. MI is not the propagated flexibility of a cloud service!
Mark Radcliffe commented
Scale operations seem to take MUCH longer now, I have waited 2+ hours to get a scale change. Why is this?