Can anyone share some lessons learned or best practices regarding SPS
2003 Alerts and index crawling configuration?

We are seeing issues where immediate alerts are not being generated
where they are expected, index crawls are failing to terminate, full
index crawls are triggering the generation of numerous old alerts
(i.e. discovering content and areas which were added months ago).

Are there best practices for:
how often to do full vs. incremental crawls?
how often to check for immediate alerts?

Are there ways to troubleshoot and/or monitor alert activity behind
the scenes so we can identify and intercept old alerts before they
are emailed to end users?

Is there a way to find out which users have alerts set to which SPS
areas and/or objects?