feat(eks/cijio-agents2) add a new 'applications' node group aimed at replacing the initial tiny_linux, to host ACP, datadog, and all technical services

[dduportal]

Ref. jenkins-infra/helpdesk#4319 (comment)

This PR adds a new node group "applications" aimed at replacing tiny_linux (which will be removed on a subsequent PR).
It also changes the following elements:

• Cleanup the moved.tf file from Chore: enforce name conventions #69 (now we renamed resources to enforce convention)
• Force the tinylinux node pool to only 1 node (to avoid autoscaler to sclae it up when draining it)
• Export the node labels and taints to the JSON output to be tracked by updatecli in kubernetes-management (for datadog and ACP at least).

Note:

• The size of the node pool is set to t4g.xlargewhich is a 4vCPUs, 16 Gb Graviton (arm64) machine
  • ACP requires 4 vCPUs and 8 Gb today as per https://github.com/jenkins-infra/kubernetes-management/blob/10eea8d9b7593a23f8b4a4ef914796de7f33c18d/config/artifact-caching-proxy_azure-aks.yaml#L36-L39
    • Note: we might remove the CPU limit to avoid CPU throttling on ACP
  • Datadog cluster-agent, like cluster-autoscaler and coredns, do not have resource limits defined, but they need some space. ACP is not expected to use all of the vCPUs: let's start with this size, and we'll see how the system behaves

[smerle33]

🚀