RLS: 2022.9.21 (#71)

fixup

Request gpu=0 for non-GPU profiles (#66)

It seems there's some process level state in Kubespawner that persists
between spawn attempts. We've observed users requesting a GPU profile,
the spawn failing (thanks to hitting our GPU quota) and then subsequent
spawns failing because they're requesting `nvidia.com/gpu=1`, even for
non-GPU profiles.

To attempt to fix this, we always set a request for nvidia.com/gpu.
We explicitly set it to 0 for non-gpu profiles.

Bump prod images (#65)

Try bumping min-count (#61)

Fixed cull level

Update environment images (#57)

Update k8s-lifecycle-hook-post-start.sh

Update kbatch-proxy-values.yaml (#51)

Bump leafmap (#46)

* Bump leafmap