We plan to do some significant work on the pollers in 2008. We'll be sure to include those idea during that work.
Here is another feature request for the Hot-Standby poller.
HS poller should be in the state to assume polling immediately upon a primary poller failure (i.e., takeover timer threshold reached). Today, upon detecting a primary poller failure, it takes several minutes to ramp up (read into memory the db) before the HS poller assumes polling.
The same goes for the hotfixes,
In V8.1 some hotfixes stop & restart the polling services, while other's do not.
I prefer to stop the services myself, install all the hotfixes and then restart the services.
Nothing like bouncing the services several times in a minute and sending our 100's of false alerts to everyone..
I normally keep the Advanced Alerting service disabled for 15 minutes, until things have stabilized after a restart.