Hi!
I have bazillions of incoming (rejected) attempts to connect to my
SMTP server, and I'm interested in separating out those that seem
to come in huge bunches (e.g., the one from yesterday that ran for
about 10 hours and sent over 4100 attempts), and graphing them so
I can see the spacing and/or distribution in time.
I can figure out some simple scripting to turn the maillog entries
into times since the epoch, or other formats, if needed, but I have
no experience with the various graphing tools availabe, or even
what (or where) they are.
I'd appreciate pointers-and-or-advice, should any of you have any
such things to give.
thanks in advance!
Fred
--
-------------------------------------------------------------------------------
Under no circumstances will I ever purchase anything offered to me as
the result of an unsolicited e-mail message. Nor will I forward chain
letters, petitions, mass mailings, or virus warnings to large numbers
of others. This is my contribution to the survival of the online
community.
--Roger Ebert, December, 1996
----------------------------- The Boulder Pledge -----------------------------
Hey Fred, If you can organize your data into a spreadsheet you can use the built in graphing abilities of LibreOffice Calc. That should be sufficient for a limited use application like you are describing. I use LibreOffice Calc to make an hourly graph of my Internet upload and download speed performance. On 06/19/17 15:26, Fred Smith wrote:> Hi! > > I have bazillions of incoming (rejected) attempts to connect to my > SMTP server, and I'm interested in separating out those that seem > to come in huge bunches (e.g., the one from yesterday that ran for > about 10 hours and sent over 4100 attempts), and graphing them so > I can see the spacing and/or distribution in time. > > I can figure out some simple scripting to turn the maillog entries > into times since the epoch, or other formats, if needed, but I have > no experience with the various graphing tools availabe, or even > what (or where) they are. > > I'd appreciate pointers-and-or-advice, should any of you have any > such things to give. > > thanks in advance! > > Fred >-- _ ?v? /(_)\ ^ ^ Mark LaPierre Registered Linux user No #267004 https://linuxcounter.net/ ****
On Mon, 2017-06-19 at 15:26 -0400, Fred Smith wrote:> I can figure out some simple scripting to turn the maillog entries > into times since the epoch, or other formats, if needed, but I have > no experience with the various graphing tools availabe, or even > what (or where) they are.mrtg would be a simple option. It's designed for graphing network traffic but can be used to graph anything you like. There are examples in the contrib directory. -- Ian
On 6/19/2017 7:25 PM, Ian Mortimer wrote:> mrtg would be a simple option. It's designed for graphing network > traffic but can be used to graph anything you like. There are examples > in the contrib directory.mrtg is very obsolete, and has been replaced with rrdtool, and integrated realtime graphing systems like cacti and librenms -- john r pierce, recycling bits in santa cruz
> > I can figure out some simple scripting to turn the maillog entries > into times since the epoch, or other formats, if needed, but I have > no experience with the various graphing tools availabe, or even > what (or where) they are. >If you are talking about scripting graphs, then Gnuplot is what you need - it's in the core centos distro I think. If you want a dedicated gui graphing application, then you can't go far wrong with Grace - it's in EPEL. I've used it for producing publication quality plots but it's equally capable of dealing with everyday graphing. The native save file format is simple and plain text, so it's quite easy to generate the files using a script which can then be read straight in to grace and will have correct axes, lables, etc. P.
On 6/20/2017 1:27 AM, Pete Biggs wrote:>> I can figure out some simple scripting to turn the maillog entries >> into times since the epoch, or other formats, if needed, but I have >> no experience with the various graphing tools availabe, or even >> what (or where) they are. >> > If you are talking about scripting graphs, then Gnuplot is what you > need - it's in the core centos distro I think. > > If you want a dedicated gui graphing application, then you can't go far > wrong with Grace - it's in EPEL. I've used it for producing publication > quality plots but it's equally capable of dealing with everyday > graphing. The native save file format is simple and plain text, so it's > quite easy to generate the files using a script which can then be read > straight in to grace and will have correct axes, lables, etc.another tool i've used for creating various sorts of graphs is Gnu Octave. this is a matlab replacement, and if you can import a bunch of numbers, you can graph them 8 ways from sideways. -- john r pierce, recycling bits in santa cruz
Am 19.06.2017 um 21:26 schrieb Fred Smith:> Hi! > > I have bazillions of incoming (rejected) attempts to connect to my > SMTP server, and I'm interested in separating out those that seem > to come in huge bunches (e.g., the one from yesterday that ran for > about 10 hours and sent over 4100 attempts), and graphing them so > I can see the spacing and/or distribution in time. > > I can figure out some simple scripting to turn the maillog entries > into times since the epoch, or other formats, if needed, but I have > no experience with the various graphing tools availabe, or even > what (or where) they are. > > I'd appreciate pointers-and-or-advice, should any of you have any > such things to give. > > thanks in advance! > > FredOut of the box: mailgraph https://mailgraph.schweikert.ch/ https://github.com/schweikert/mailgraph There are forks which implement newer anti-spam implementations like Postfix's postscreen. See an example here https://mailgraph.piratenpartei-bayern.de/ Something more generic but powerful: Prometheus with Grafana It works with a time series database in the background and is the de-facto standard in DevOps environments. With a Google image search for Grafana you find a lot of examples how it looks like. If the application does not provide the necessary metrics itself for Prometheus then an exporter is required. You find some for Postfix. Prometheus will scrape the exported metrics and within Grafana you can easily add nice and useful graphs into your dashboards. Alexander