Give k8s-infra-prow-build cluster time to scale-up

We have k8s-infra-prow-build etup to autoscale if it doen't have
enough capacity. Anecdotally I saw it take 5sec to decide that
it needed to scale, but another ~2min for a node to come online.

Let's assume it could take longer, so wait up to 5min before calling
an unscheduled pod safe to delete.

config/prow/config.yaml

@@ -5,7 +5,7 @@ plank:
   job_url_prefix_config:
     '*': https://prow.k8s.io/view/
   pod_pending_timeout: 15m
-  pod_unscheduled_timeout: 1m
+  pod_unscheduled_timeout: 5m
   default_decoration_configs:
     '*':
       timeout: 2h