Robert Brooks wrote:
Mark Martinec wrote:
I mean something like this...
Spam Score -5 0 3 5 10 15
Percentile5.89% 12.59% 19.43% 24.81% 35.02% 50.37%
Seems like you want a frequency distribution: divide a score range
into arbitrary buckets,
MrC wrote:
Robert Brooks wrote:
[..]
Oct 5 09:01:06 mailhost amavis[25870]: (25870-17) Blocked SPAM,
[84.79.70.165] [84.79.70.165] [EMAIL PROTECTED] -
[EMAIL PROTECTED], mail_id: HFW3sN8D+u30, Hits: 20.862, size: 719,
11887 ms
[..]
Without data in the log that supports such an
MrC wrote:
Robert Brooks wrote:
[..] (discussing postfix-logwatch log analyser)
whilst I'm being a pain the Postfix reject figures don't show
rejections from before-queue amavis. Unfortunately Postfix doesn't
seem to log this well :(
I don't use a before-queue setup, and don't know what
Robert Brooks wrote:
MrC wrote:
Robert Brooks wrote:
[..] (discussing postfix-logwatch log analyser)
whilst I'm being a pain the Postfix reject figures don't show
rejections from before-queue amavis. Unfortunately Postfix doesn't
seem to log this well :(
I don't use a before-queue
MrC wrote:
Mark Martinec wrote:
I mean something like this...
Spam Score -5 0 3 5 10 15
Percentile5.89% 12.59% 19.43% 24.81% 35.02% 50.37%
Seems like you want a frequency distribution: divide a score range
into arbitrary buckets, fill them by
Robert Brooks wrote:
MrC wrote:
Mark Martinec wrote:
I mean something like this...
Spam Score -5 0 3 5 10 15
Percentile5.89% 12.59% 19.43% 24.81% 35.02% 50.37%
Seems like you want a frequency distribution: divide a score range
into arbitrary
MrC wrote:
Robert Brooks wrote:
MrC wrote:
I see. Its easy enough to implement. I had originally not seen much
value in those numbers, but will be happy to add the feature if it is
useful.
seeing the overall performance of amavis is useful to me, especially
as I use amavis
Robert Brooks wrote:
MrC wrote:
Robert Brooks wrote:
MrC wrote:
I mean something like this...
Spam Score -5 0 3 5 10 15
I mean something like this...
Spam Score -5 0 3 5 10 15
Percentile5.89% 12.59% 19.43% 24.81% 35.02% 50.37%
Seems like you want a frequency distribution: divide a score range
into arbitrary buckets, fill them by counts of messages falling
Mark Martinec wrote:
I mean something like this...
Spam Score -5 0 3 5 10 15
Percentile5.89% 12.59% 19.43% 24.81% 35.02% 50.37%
Seems like you want a frequency distribution: divide a score range
into arbitrary buckets, fill them by counts of
Mark Martinec wrote:
I mean something like this...
Spam Score -5 0 3 5 10 15
Percentile5.89% 12.59% 19.43% 24.81% 35.02% 50.37%
Seems like you want a frequency distribution: divide a score range
into arbitrary buckets, fill them by counts of
MrC wrote:
I see. Its easy enough to implement. I had originally not seen much
value in those numbers, but will be happy to add the feature if it is
useful.
seeing the overall performance of amavis is useful to me, especially as
I use amavis before-queue.
also, spam score percentiles. It
Robert Brooks wrote:
MrC wrote:
I see. Its easy enough to implement. I had originally not seen much
value in those numbers, but will be happy to add the feature if it is
useful.
seeing the overall performance of amavis is useful to me, especially as
I use amavis before-queue.
Ok,
Robert Brooks wrote:
MrC wrote:
Benedict - feel free to drop me a line if you have questions or
improvement requests. I have just made some recent experimental
changes in the summary presentation, and am looking for feedback on
that in particular.
one thing I'd like to see is the
MrC wrote:
Hi Rob,
Is your request different from the timings percentiles report that
currently exists ? Example at the end of:
http://www.mikecappella.com/logwatch/example-amavis-detail10
It is obtained with detail level = 5.
this is what I'm thinking about, but as well as the
Robert Brooks wrote:
MrC wrote:
Hi Rob,
Is your request different from the timings percentiles report that
currently exists ? Example at the end of:
http://www.mikecappella.com/logwatch/example-amavis-detail10
It is obtained with detail level = 5.
this is what I'm thinking about,
Yesterday MrC said:
Mark Martinec wrote:
Seems the only one that is actively maintained is the logwatch:
http://www.mikecappella.com/logwatch/
Benedict - feel free to drop me a line if you have questions or
improvement requests. I have just made some recent experimental changes
in
+ or ($p1 =~ /^AM.PDP /) # this appears to be always two spaces
^^^
or ($p1 =~ /email.txt no longer exists, can't re-use it/)
^^^
Don't forget to quote literal dots: \. is a dot, a . matches any character.
Mark
Curtis Doty wrote:
Here's a couple of tweaks needed to quiet down the unmatched noise a bit:
--- amavis-logwatch.orig2007-09-17 21:07:47.0 -0700
+++ amavis-logwatch2007-09-21 09:43:46.0 -0700
@@ -458,6 +458,10 @@
or ($p1 =~ /^updating snmp variables/)
has anyone got a recomendation for a log file analyser for both Amavis and
Postfix.
I have got some reasonable ones for each but could do with one that works with
both.
Kind regards
Benedict White
-
This SF.net email is
Benedict White wrote:
has anyone got a recomendation for a log file analyser for both Amavis and
Postfix.
I have got some reasonable ones for each but could do with one that works
with both.
mailgraph
Kind regards
Benedict White
Benedict,
has anyone got a recomendation for a log file analyser for both Amavis and
Postfix. I have got some reasonable ones for each but could do with one
that works with both.
Seems the only one that is actively maintained is the logwatch:
http://www.mikecappella.com/logwatch/
Mark
Benedict White wrote:
has anyone got a recomendation for a log file analyser for both Amavis and
Postfix.
I have got some reasonable ones for each but could do with one that works
with both.
Kind regards
Benedict White
Mark Martinec wrote:
Benedict,
has anyone got a
23 matches
Mail list logo