Re: little off topic monitoring question
Shane, Thanks Mark, We are using Amavisd to call SpamAssassin. Cacti looks great, we hope to implement Cacti soon. Thanks for the info, and many thanks for Amavisd. Btw, for more info search the release notes (the feature was introduced in 2.6.4) for the following text: - newly supplied with the package is a program amavisd-snmp-subagent, acting as an SNMP AgentX, exporting amavisd statistical counters database (snmp.db) as well as a child process status database (nanny.db) to a SNMP daemon supporting the AgentX protocol (RFC 2741), such a NET-SNMP. [...] Mark
Re: little off topic monitoring question
Shane, We would like to start monitoring our two smtp servers. They are fairly busy boxes, maybe 100,000 messages a day, give or take several thousand. They of course run Spamassassin, Postfix is also used. We use MRTG to monitor internal servers and switches, and would really like something with a similar graph. If you are using amavisd to call SpamAssassin (instead of spamd), then there is a huge number of SNMP counters and gauges to monitor the operation in real time, including SNMP gauges on current Postfix queue sizes. See: http://www.ijs.si/software/amavisd/AMAVIS-MIB.txt for a list of available OIDs. It involves a lightweight daemon amavisd-snmp-subagent, which collects statistics from amavisd statistics database in almost- real time, and presents itself as na AgentX to a Net-SNMP daemon, which deals with the SNMP protocol. We are using Cacti for monitoring a selection of these stats. No need to parse log files. See also http://www.amavis.org/amavis-2011.pdf slide 90 and onwards. Mark
RE: little off topic monitoring question
Thanks Mark, We are using Amavisd to call SpamAssassin. Cacti looks great, we hope to implement Cacti soon. Thanks for the info, and many thanks for Amavisd. Shane -Original Message- From: Mark Martinec [mailto:mark.martinec...@ijs.si] Sent: Wednesday, July 20, 2011 7:25 AM To: users@spamassassin.apache.org Subject: Re: little off topic monitoring question Shane, We would like to start monitoring our two smtp servers. They are fairly busy boxes, maybe 100,000 messages a day, give or take several thousand. They of course run Spamassassin, Postfix is also used. We use MRTG to monitor internal servers and switches, and would really like something with a similar graph. If you are using amavisd to call SpamAssassin (instead of spamd), then there is a huge number of SNMP counters and gauges to monitor the operation in real time, including SNMP gauges on current Postfix queue sizes. See: http://www.ijs.si/software/amavisd/AMAVIS-MIB.txt for a list of available OIDs. It involves a lightweight daemon amavisd-snmp-subagent, which collects statistics from amavisd statistics database in almost- real time, and presents itself as na AgentX to a Net-SNMP daemon, which deals with the SNMP protocol. We are using Cacti for monitoring a selection of these stats. No need to parse log files. See also http://www.amavis.org/amavis-2011.pdf slide 90 and onwards. Mark
Re: little off topic monitoring question
On 2011-07-19 14:50, Thomas Mullins wrote: We would like to start monitoring our two smtp servers. They are fairly busy boxes, maybe 100,000 messages a day, give or take several thousand. They of course run Spamassassin, Postfix is also used. We use MRTG to monitor internal servers and switches, and would really like something with a similar graph. Thanks Shane g.co sez http://mailgraph.schweikert.ch/
Re: little off topic monitoring question
On Tue, July 19, 2011 14:50, Thomas Mullins wrote: We would like to start monitoring our two smtp servers. They are fairly busy boxes, maybe 100,000 messages a day, give or take several thousand. They of course run Spamassassin, Postfix is also used. We use MRTG to monitor internal servers and switches, and would really like something with a similar graph. I like this: http://mailgraph.schweikert.ch/
Re: little off topic monitoring question
On 19.07.2011, at 14:50, Thomas Mullins wrote: We would like to start monitoring our two smtp servers. They are fairly busy boxes, maybe 100,000 messages a day, give or take several thousand. They of course run Spamassassin, Postfix is also used. We use MRTG to monitor internal servers and switches, and would really like something with a similar graph. Thanks Shane pflogsumm is also interesting and also logwatch http://logreporters.sourceforge.net/
Re: little off topic monitoring question
We would like to start monitoring our two smtp servers. They are fairly bu= sy boxes, maybe 100,000 messages a day, give or take several thousand. The= y of course run Spamassassin, Postfix is also used. We use MRTG to monitor= internal servers and switches, and would really like something with a simi= lar graph. Thanks Shane Hi, if this is about monitoring (rather than just collecting statistics): At some time I had set up a system that would simply send mail via the server and receive it back. If that did not happen (within reasonable time), something might be wrong. Basically the system was 4 separate components (smtp, pop/imap, spam filtering, and a database holding everything together). It occasionally happened that things got stuck, so one could, perhaps, successfully connect to smtp ... but smtp accumulated a pile of messages that SA did not want to process Wolfgang
Re: little off topic monitoring question
On 7/19/2011 8:50 AM, Thomas Mullins wrote: We would like to start monitoring our two smtp servers. They are fairly busy boxes, maybe 100,000 messages a day, give or take several thousand. They of course run Spamassassin, Postfix is also used. We use MRTG to monitor internal servers and switches, and would really like something with a similar graph. I've been very pleased with www.websitepulse.com They do a round trip smtp-send/pop-retrieval. I get text messaged if this ever fails. I also used use them for http-checking my webmail. -- Rob McEwen http://dnsbl.invaluement.com/ r...@invaluement.com +1 (478) 475-9032