Description
We currently don't orchestrate fail-over for a Service, if the node failed over is the last node the Service is hosted on (I am not sure if this was intended). Opening this MB to explore if we can always explicitly call orchestrate_service_failover/2 here -> https://github.com/couchbase/ns_server/blob/master/src/service_janitor.erl#L200.
See CBSE-13766 for more details on why cbas needs this.