--

Interesting. I read about the Autopilot a few days ago. Wasn’t sure about what’s the added value.

Honestly, I still don’t.

Ok, so it can prevent wasting resources by downscaling against settings. But what’s the use of that if by default an Autopilot cluster costs more anyway?

Also, even a classic cluster can autoscale its nodes on a per pool basis to accommodate load.

And there’s HPA to help as well as VPA (soon) so in terms of load management I can see how Autopilot might be more efficient at handling this overall (scaling up as well as down) but is it worth double the cost? Not sure.

--

--