* Issue #69 - Auto-sync option in application CRD instance

* Fix possible NPE error in controller

* Correctly handle situation when cluster is temporary down/unreachable by argocd

* Handle case when argo monitor cluster which go offline and recover