We have pound setup with some spot market AWS EC2 instances as some of the backends.
Spot instances can suddenly be taken off line. It seems that it takes a while for pound to respond to events like this - according to another post several tcp connections refused or not respond for pound to recognize the server is "dead". If we set up an external monitoring of the backends and detect that some are down (or likely to go down shortly because of pricing of the spot market) is there a way to inform pound that server x is now dead as opposed to waiting for pound to discover it on it's own? Other systems like bind have facilities to send a signal to reload the config file but I did not see a facility like that in pound. TIA johnh...
