We need a documented and supported API to alert administrators

Memcached needs a way to alert the administrators for various reasons (configuration problems, consistency issues detected, malicious clients detected, resource leaks etc). As of right now we can just put this information in memcacheds own logfiles (or write to stderr and potentially get it in ns_servers logfiles unless we send too much data causing it to get dropped) and cross our fingers that an operator detects it.

There is already a log tab in our UI where all components SHOULD be able to put information in human readable form so that the operators could resolve problems. We need an API for the other components in the system to put information here. In addition to that we need an API to send alerts to the administrators that pops up like the "disk full" events.