> All sendmail statistics are generated by separate programs which process
> the system log file or the sendmail stats file.
>
> There is/was a flurry of interest in this stuff years ago (circa 1993) on
> the perl list. And in fact, Tom Christiansen himself wrote at least one
> version.
>
> This is a concept/request which florishes from time to time. The problem is
> that what you are asking for is rediculous unless you have a very small
> system with very few users who contact very few outsiders. It is like all
> "accounting" functions -- if you have a body or two to devote to writing
> the programs and maintaining the data collected, it can be useful, but it
> is typical "graduate student work."
>
> Basically, the report you are asking for is pretty much as worthless as the
> daily and weekly reports generated by netnews.
>
> The "top 10" information is sometimes informative, but the bulk of the
> information is far less meaningful than people realize.
>
> For one thing, there are a variety of issues which effect things like the
> "Time frame breakdown" including the granularity of the logging in the
> first place.

I agree that it isn't as easy to generate meaningful statistics as it first
appears.  But that doesn't mean that no statistics would be meaningful.  I'm
thinking that would would be interesting is

(1)  Overall volume and the trends in that.

(2)  Delivery time for the vast majority of messages, and the trends in
that.  Distinguishing between internal and external traffic sounds useful.

If we were to see delivery time trending up, that might be something we
should look into before it gets really bad and people start complaining.

Things like overall volume can be useful to present to upper management
(to justify a bigger machine, for example).

Could you further explain what you mean by problems introduced by the
granularity of the logging?

The system I am most interested in is the one that we have dedicated to
LISTSERV.