Network performance monitoring with GTM?

I have two sites - lets call them A and B. Each site has a single GTM (VE) and two LTMs (BigIP 6900). Site A is on the East Coast of the US, and Site B is on the West Coast.

 

Site B occasionally suffers a network performance degradation (apparently the ISP is being subjected to a DDoS attack, which results in increased ping times up to 700ms, and packet loss of up to 50%).

 

Neither of the GTMs have so far successfully "caught" this connectivity degradation and continued sending traffic to the Site B.

 

The question is how do I prevent this from happening? I need to have Site A monitor Site B's connectivity and remove Site B from the loop if the response time exceeds a given amount in milliseconds, or if packet loss occurs. An obvious solution would be a specific monitor configuration, where GTM at Site A pings GTM at Site B, but the monitors can only have timeout values in seconds. But when I create my own Gateway_ICMP monitor with 1-second probe timeout - the monitored services start "flip-flopping", going up and down every second or so...

 

I've tried to find information on Prober Pools, but everything I could find is extremely vague and generic :(

 

Any help would be GREATLY appreciated!