Make sure we have 2 worker nodes in advance (#4787)

Looking at the tekton task throttled by node, we can see that one node
in advance is helping but not enough to get rid of task pending due to
lack of nodes.

Signed-off-by: Hugo Ares <[email protected]>

configs/proactive-scaler/base/deployments.yaml

@@ -5,7 +5,7 @@ metadata:
   name: m6a-4xlarge
   namespace: proactive-scaler
 spec:
-  replicas: 1
+  replicas: 2
   selector:
     matchLabels:
       run: m6a-4xlarge