Request half of a CPU per user session

This value was changed, as the puppet nodes are being taken out of
service and so, there will be a higher number of user sessions running
in the SWAN kubernetes cluster.

Therefore, taking into account the maximum number of CPUs available, it
was decided to request half of a CPU per user session, to allow for
more user sessions to be scheduled.

This level of overcommiting is fine, under the assumption most users will
be idle for the majority of the time and will only require more
CPU resources when executing notebook cells.

swan-cern/values.yaml

@@ -71,7 +71,7 @@ swan:
   jupyterhub:
     singleuser:
       cpu:
-        guarantee: 1
+        guarantee: 0.5
     scheduling:
       userPods:
         nodeAffinity: