Add create-monitor=true to could.conf to solve LB issues.

Created by: garloff

See issue #173 (closed): When the LB tries to talk to all 3 nodes in round robin mode, it always typically fails twice before finding the one member where the nginx L7 LB accepts the connections. Having a (TCP) health monitor makes the LB aware of which members respond and which ones do not.

This can be achieved by the global setting create-monitor=true in OCCM's cloud.conf configuration file. The only question is: Why does it default to false currently?

Signed-off-by: Kurt Garloff kurt@garloff.de