Nimbostratus
Cirrostratus
By default, LTM monitors and load balanced based on the pool member IP:port. So if each service is running on a separate TCP port, you could just configure the IP:port combinations in a pool and use a monitor to check the service is answering before using the pool member. You can use the 'action on service down' property on the pool to specify what LTM should do for existing connections if a pool member is marked down.
Which protocol(s)/applications is this for? If it's a generic TCP service, you could use a TCP monitor or a TCP half open monitor. If it's HTTP or HTTPS you could configure one of those monitors.
Aaron