On Mon, 09 Aug 1999 at 10:47:53 -0400, Cris Daniluk wrote:
> We've been tryin to do some more reliable and dependable benchmarking of
> qmail in our experimentation, but it is very difficult. One of the "nice
> things" (one of?) about ms smtp server is the fact that nt performance
> monitor can keep track of it. We can see how many messages per second are
> being sent, how many concurrent connections, etc. While it is easy to
> monitor concurrent connections by a good old ps auxww |grep qmail-remote
> |wc -l, it is not exactly well represented, nor scientific. Messages per
> second can really only be done by monitoring qstat and a stop watch, or by
> sending to a pool of ms smtp servers and monitoring them. It would be truly
> wonderful if we could get more concrete and scientific results of all of
> this, because I think it would be an excellent way to demonstrate the power
> of qmail, and I think it would be a wonderful tribute to everyone who has
> put their time into the project. They all deserve a lot of credit, and this
> seems to be the best way to give them that due credit. Are there any
> available benchmarking tools? If not, I think it would be worthwhile to

Have you tried qmailanalog? 
http://pobox.com/~djb/qmailanalog.html

It gives _among others_:

the time spent by a message in the queue,
the latency for a delivery attempt,
latency for a successful delivery to one recipient,
the first pct% of successful deliveries all happened within doneby secs,
the number of messages sent by the sender,
the number of bytes sent by the sender,
the number of bytes successfully received from the sender,
the number of bytes from the sender, weighted by recipient,
the number of delivery attempts (success, failure, deferral)
the number of recipients (success plus failure),
the number of bytes successfully delivered to the recipient,
the number of messages sent to the recipient (success plus failure),
the number of bytes successfully delivered to the given host,
the number of messages sent to the given host (success plus failure)

-- 
 Tomasz Papszun   SysAdm @ TP S.A. Lodz, Poland  | And it's only
 [EMAIL PROTECTED]   http://www.lodz.tpsa.pl/   | ones and zeros.

Reply via email to