Environment sustainability and autoscaling
My name is Tom Kerkhove and I'm one of the maintainers of Kubernetes Event-driven Autoscaling (KEDA) that allows end-users to efficiently scale workloads.
We have a strong focus on scaling applications with low footprint and have started a thread in Slack earlier this week.
While we are not committing to anything, I'm very curious to hear feedback from sustainability experts to see how we can help as an autoscaler to reduce the impact on our environment.
One idea in the thread is to introduce a carbon scaler that will only allow you to scale out if the carbon impact is not too high.
Another idea I personally have is to see if we can help with the scheduling process to go with spot instances when available over more beefy ones to reduce compute waste.
PS: For those interested, there was a very nice talk at KubeCon NA last year that talked about using KEDA to reduce environmental impact.
So in short - How can we help reduce the impact on the environment as an autoscaler? Let your wildest ideas become a reality (if we can make it happen).