self-hosted/docs: Use reasonable node min, max, defaults for proof of value reference architecture

[adrienthebo]

Is your feature request related to a problem? Please describe

At present the GKE proof of value cluster architecture uses the following values

Setting	Default
Starting node count	1
Min node count	1
Max node count	50

When working on the EKS proof of value documentation it was discovered that a single node cluster doesn't have enough CPU to create a workspace as the Gitpod services request enough CPU to max out a single node; at least two nodes are needed. The draft EKS proof of value documentation currently uses a max node value of 4, which is divergent from our pre-existing defaults.

This issue is exacerbated by #12410 where EKS clusters don't yet autoscale, so the default configuration will not work. Even if autoscaling is operable, the proof of value guides should aim to create a cluster what will be able to immediately spin up a workspace without scaling up to provide the best possible experience.

Describe the behaviour you'd like

The proof of value reference architectures should allocate enough resources to create at least one workspace without needing to scale up.

This will likely look like the following but more testing is needed:

Setting	Default
Starting node count	2
Min node count	1
Max node count	4

[lucasvaltl]

@adrienthebo is this still relevant now that we are using 16 core nodes as the default for the PoV ref arch?

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.