[CX] analytics-docker not reliably configuring cluster w/ desired services

For about half of the Analytics team members (myself, Murtadha, Xikui) the analytics-docker image starts up with only kv provisioned.  Conjecturing that perhaps 15s isn't long enough on all machines for the server to start, can we update this to wait for the server to come up before issuing configuration commands?  Our tests wait for a 200 on :8091, followed by a 2s sleep before proceeding to configure, which has proven reliable for us.

I think it would also be good to capture the output (curl -v) and present it in the event curl returns a non-zero error code, to assist in future debugging / customer error reporting.