I am just playing with ideas for processing exim logs.
Our system is a little more complex than most in that we have several
parallel systems running together and I would like to be able to get stats
for them as a whole as well as some information about the component
systems.
Hence I am thinking about making the basic log parsing and the stats
generation components of the log processing separate so that the log could
be reduced in size substantially. The stats generation side could then
take in files from n servers to give an aggregate, or one of n stats
generators could be used to generate stats in text, html, klingon etc.
The problem is really defining the format of the intermediate file.
One easy method is to pretty much dump the raw data structures after the
parse phase... this is actually very easy using the Data Dumper package,
but has the disadvantage that the data written is as a perl script and you
eval it back in, which is a tad dangerous!
As I said I am just playing with ideas right now so good thoughts, ideas
about what is really wanted from a stats set for email etc would all be
useful to me...
Nigel.
--
[ Nigel.Metheringham@??? - Systems Software Engineer ]
[ Tel : +44 113 251 6012 Fax : +44 113 224 0003 ]
[ Friends don't let friends use sendmail! ]
