Atelet should only run on nodes where ateoms are running.

Today, atelet is deployed as a DaemonSet, which means it runs on all nodes in the cluster, taking into account taints and tolerations.

Ateom pods are managed as Deployments today, though we may eventually switch to a custom workload controller.

We should find a solution that places atelet pods only on nodes where ateoms are or will be running.

This could look like:
* Manage atelets from the same workload controller as ateoms, so that it can place atelet pods on the correct nodes
* Come up with a convention for tainting certain nodes as only for ateoms, not regular workloads.