[squid-users] interpreting results of cachemgr.cgi

Are there any detailed notes on how to use the information presented via
cachemgr.cgi to help diagnose performance and tune the cache? I'm
impressed by the information that is available but have a hard time
using the data as any useful output. I'd like to know things like how to
get a correct number of redirector instances started (e.g., is 20
squidGuard processes overkill for my proxy?), file descriptors to
allocate, disk/memory space/cache configuration/etc.

TIA

-- 
Darren Spruell