`WasmCloudHostConfig` should have a status that indicates if it can't scale all the way

[thomastaylor312]

I was running in GKE in autopilot mode and only 1 node had spun up. Even though I had requested 2 replicas in daemonset mode, it only gave me one (which makes sense). However, we should probably have some sort of indicator for when nodes are full or otherwise unavailable that the number of hosts desired is not possible.

This is really low priority, was just pointing it out since I discovered it

[f4z3r]

Hi @thomastaylor312 I am afraid that is not how daemonsets work. Kubernetes Daemonsets will always schedule a pod on all eligible nodes based on selectors. Thus what you enter as host_replicas in the WasmCloudHostConfig becomes irrelevant and will not be used by the operator. This is not very nice in terms of configuration I have to admit though, as you would expect the field to be respected. I am working on implementing a better status for the WasmCloudHostConfig so that you can see how many replicas are currently deployed, regardless of whether you are running in daemonset mode or not. Thus you can validate whether the deployment of the hosts has been performed already or not. Nonetheless, this will not fix that if you, for instance, have host_replicas set to 2, it will deploy only a single host if only one eligible node is available, and might deploy 4 hosts if 4 nodes become eligible.