Industry Acceptable % of Timeout failures (502s)

["Jignesh Badani" <jbadani@xxxxxxxx>]

Hello, is there an industry standard for
acceptable % of Timeout related failures on one's website ?

We have a server farm of 3 Apache RP
proxying almost everything to the back-end content and app farms of about
40-50 servers. The user base is small but fairly active (about 2.5mil hits
per day & 17GB-20GB traffic served from all RPs combined)

We have a current Timeout related 502
messages of about 0.3% of the total successful hits (http 200).  Is
that on the high side or acceptable? If there is an website about the same,
please point me to it.

The reason I ask is that we are very
short handed and we cannot go really chasing the App or the content folks
for every 502s getting logged.

regards
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
- - - 
JB