Kuberenetes autoscale is super hacky

[Dougley]

I haven't found a reliable way of detecting what scale a workload runs at, and what index that pod would have inside the workload, aside from statically passing the scale of the workload as an environment var and (ab)using statefulsets to get the index of the pod by parsing the hostname.

Obviously this is extremely inefficient and fragile, and I would like something more robust in the future