[Kubernetes CI] Allow pods to be scheduled in the default node pool

Regularly, the different teams are suffering from pods never being scheduled: Platform, UI, AI.
----------------------------------------------------
Because of the nodes-startup Daemonset installed in the admin-nodes-startup namespace, the nodes of the default pool "pool-2" are tainted with:

 taints:
 - effect: NoSchedule
   key: dedicated
   value: nodes-startup

Theoretically, this taint shoud disappear at some point, but it seems not to.
Consequently, no pod can get scheduled in this node pool, unless it has the "dedicated: nodes-startup" toleration.
Thus, a pod without any specific toleration that would allow it to be scheduled on a node pool dedicated to a team will never be scheduled.
Typical symptom:

Normal   NotTriggerScaleUp  27s                cluster-autoscaler  pod didn't trigger scale-up: 1 node(s) had taint {team: nos}, that the pod didn't tolerate, 1 node(s) had taint {team: platform}, that the pod didn't tolerate, 1 node(s) had taint {team: ui}, that the pod didn't tolerate, 1 node(s) had taint {team: napps}, that the pod didn't tolerate, 1 node(s) had taint {dedicated: nodes-startup}, that the pod didn't tolerate, 1 max node group size reached
  Warning  FailedScheduling   16s (x4 over 29s)  default-scheduler   0/28 nodes are available: 1 node(s) had taint {team: ui}, that the pod didn't tolerate, 10 node(s) had taint {team: ai}, that the pod didn't tolerate, 5 node(s) had taint {team: napps}, that the pod didn't tolerate, 6 node(s) had taint {dedicated: nodes-startup}, that the pod didn't tolerate, 6 node(s) had taint {team: platform}, that the pod didn't tolerate.

List of possible tolerations for the team dedicated node pools:

• team: platform
• team: napps
• team: ui
• team: nos
• team: ai

Though the recommendation is to always set tolerations/nodeSelector on the Pods/Deployments/Statefulsets/... to get attached to a team dedicated node pool, we should allow pods to be scheduled in this "pool-2" default node pool.

Simple solution: remove the "nodes-startup" Daemonset. It should not be needed.
If problematic we can create it back.