Re: alerts functionality
On Fri, 19 Nov 2004, Joubin Moshrefzadeh wrote: > host1 goes down - 1 alert sent > then host2 goes down - 2 alerts sent > then host3 goes down - 3 alerts sent > etc... > > so total alerts sent is 1+2+3...+10? > > is the latter correct? I've only tested it up to two hosts going down > consecutively :) it's correct depending on how you configure mon. this is the default behavior, but you can change it. i noticed the man page needed some updating, so i did so and check in the changes to the cvs tree on the mon-1-0-0pre1 branch. the part which affects this behavior is the "alertevery" parameters. here's a summary: ALERT DECISION LOGIC Upon a non-zero or zero exit status, the associated alert or upalert program (respectively) is started, pending the following conditions: If an alert for a specific service is disabled, do not send an alert. If dep_behavior is set to 'a', and a parent dependency is failing, then suppress the alert. If the alert has previously been acknowledged, do not send the alert, unless it is an upalert. If an alert is not within the specified period, record the failure via syslog(3) and do not send an alert. If the failure does not fall within a defined period, do not send an alert. No upalerts are sent without corresponding down alerts, unless no_comp_alerts is defined in the period section. An upalert will only be sent if the previous state is a failure. If an alert was already sent within the last alertevery interval and the monitor has continued to report a nonzero exit status for a time period less than that interval, do not send another alert, unless the summary output from the most recent monitor process differs from the previous. Other- wise, send an alert using each alert program listed for that period. The observe_detail argument to alertevery affects this behavior by observing the changes in the detail part of the output in addition to the summary line. If a monitor has successive failures and the summary output changes in each of them, alertevery will not suppress multiple consecutive alerts. The reasoning is that if the summary output changes, then a significant event occurred and the user should be alerted. The "ignore_summary" option will suppress all successive alerts while the service continues to fail, even if the summary output changes. If the "strict" alertevery option is used, then behave the same as if "ignore_summary" was set, but do not reset the alertevery timer when the monitor exits with a zero status. For example, "alertevery 24h strict" will only send out an alert once every 24 hours, regardless of whether the monitor output changes, or if the ser- vice stops and then starts failing. ... alertevery timeval [observe_detail | ignore_summary | strict ] The alertevery keyword (within a period definition) takes the same type of argument as the interval variable, and limits the number of times an alert is sent when the service continues to fail. For example, if the interval is "1h", then the alerts in the period section will only be triggered once every hour as the service continues to fail. The alertevery interval timer will be reset if the monitor stops exiting with a nonzero exit status (i.e. it reports a success). If the alertevery keyword is omit- ted in a period entry, an alert will be sent out every time a failure is detected. By default, if the summary output of two successive failures changes, then the alertevery interval is overridden, and an alert will be sent. The "ignore_summary" argument suppresses this behavior. Ifthestring "observe_detail" is the last argument, then both the summary and detail output lines will be considered when comparing the output of successive failures. If the string "strict" is the last argument, then the output of the monitor or the state change of the service will have no effect on when alerts are sent. That is, "alertevery 24h strict" will send only one alert every 24 hours, no matter what. Please refer to the ALERT DECISION LOGIC section for a detailed explanation of how alerts are suppressed. ___ mon mailing list [EMAIL PROTECTED] http://linux.kernel.org/mailman/listinfo/mon
Re: alerts functionality
--On Monday, November 22, 2004 1:45 PM -0500 Jim Trocki <[EMAIL PROTECTED]> wrote: so total alerts sent is 1+2+3...+10? is the latter correct? I've only tested it up to two hosts going down consecutively :) it's correct depending on how you configure mon. this is the default behavior, but you can change it. Also, it should be pointed out that this is entirely dependent on the behavior of the monitor script. If the script outputs a different summary, then Mon will alert again (unless configured not to). Most scripts output the list of failing hosts as the summary, but not all. -David David Nolan<*>[EMAIL PROTECTED] curses: May you be forced to grep the termcap of an unclean yacc while a herd of rogue emacs fsck your troff and vgrind your pathalias! ___ mon mailing list [EMAIL PROTECTED] http://linux.kernel.org/mailman/listinfo/mon
Repeat sending of upalerts
Hi ... I 've set up an a mon-server which receives traps from some client mon servers (send via remote.alert). After restarting the main mon server, all services monitored by the client mon servers are in unknown-state until I restart the client server or a service on client server fails. Is there a possibility to repeat sending upalerts to inform the main server that everything is ok and not unknown? Any other hints to get the correct status at the main mon server without restarting mon on every client? Thanks for every hint ... Regards Oliver -- Oliver Nyderle <[EMAIL PROTECTED]> ANDURAS service solutions AG Innstraße 71 - 94036 Passau - Germany Web: www.anduras.de - Tel: +49 (0)851-4 90 50-0 - Fax: +49 (0)851-4 90 50-55 Rechtsform: Aktiengesellschaft - Sitz: Passau - Amtsgericht Passau HRB 6032 Mitglieder des Vorstands: Sven Anders, Marcus Junker, Michael Schön Vorsitzender des Aufsichtsrats: Dipl. Kfm. Karlheinz Antesberger ___ mon mailing list [EMAIL PROTECTED] http://linux.kernel.org/mailman/listinfo/mon
RE: alerts functionality
As a side note to this conversation, we do our configuration a wee bit differently. We require the ability to get stats, disable services and get alerts specific to each host. Basically, we need everything to happen at the host level. In order to accomplish this, we map one host group to one host. The bonus to this configuration is that we always know specifically which service on which host is down when we receive a page. The pages we get look like $host/$server - $short_summary\n$date. If http on server a fails, we get one page plus pages every X minutes. If server b's http fails, we get one page plus pages every X minutes. The negatives, of course, are that we have a lot of perl forks happening on our monitoring servers. > -Original Message- > From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] > On Behalf Of David Nolan > Sent: Tuesday, November 23, 2004 7:31 AM > To: mon_list > Subject: Re: alerts functionality > > > > --On Monday, November 22, 2004 1:45 PM -0500 Jim Trocki > <[EMAIL PROTECTED]> wrote: > > >> so total alerts sent is 1+2+3...+10? > >> > >> is the latter correct? I've only tested it up to two hosts going down > >> consecutively :) > > > > it's correct depending on how you configure mon. this is the default > > behavior, but you can change it. > > Also, it should be pointed out that this is entirely dependent on the > behavior of the monitor script. If the script outputs a different > summary, > then Mon will alert again (unless configured not to). Most scripts output > the list of failing hosts as the summary, but not all. > > -David > > David Nolan<*>[EMAIL PROTECTED] > curses: May you be forced to grep the termcap of an unclean yacc while > a herd of rogue emacs fsck your troff and vgrind your pathalias! > > ___ > mon mailing list > [EMAIL PROTECTED] > http://linux.kernel.org/mailman/listinfo/mon ___ mon mailing list [EMAIL PROTECTED] http://linux.kernel.org/mailman/listinfo/mon
Re: Repeat sending of upalerts
--On Tuesday, November 23, 2004 5:01 PM +0100 Oliver Nyderle <[EMAIL PROTECTED]> wrote: Hi ... I 've set up an a mon-server which receives traps from some client mon servers (send via remote.alert). After restarting the main mon server, all services monitored by the client mon servers are in unknown-state until I restart the client server or a service on client server fails. Is there a possibility to repeat sending upalerts to inform the main server that everything is ok and not unknown? Any other hints to get the correct status at the main mon server without restarting mon on every client? Thanks for every hint ... I strongly recommend that you look at the current version of Mon in CVS. One of the new features is a mon config option called 'redistribute', which takes the name of an alert script and calls that alert script on *every* status update. This is to allow you to redistribute complete status information to remote mon servers, or even into other systems. We use it with a script that sends traps to other mon servers, to provide the complete status information to all our servers. I actually just committed some updates to the CVS HEAD last week, pulling up some changes that Jim has done in the Mon 1.0 cvs branch. I believe I'm about ready to tag this as mon-1.1pre and release a tarball. -David Nolan Network Software Designer Computing Services Carnegie Mellon University ___ mon mailing list [EMAIL PROTECTED] http://linux.kernel.org/mailman/listinfo/mon
Re: alerts functionality
> it's correct depending on how you configure mon. this is the default > behavior, but you can change it. > > ... Thanks for the clarification Jim. ___ mon mailing list [EMAIL PROTECTED] http://linux.kernel.org/mailman/listinfo/mon
RE: alerts functionality
--On Tuesday, November 23, 2004 8:47 AM -0800 Gary Richardson <[EMAIL PROTECTED]> wrote: The negatives, of course, are that we have a lot of perl forks happening on our monitoring servers. My servers already fork 500K times a day... That would be *so* much worse. If only we had per-host status tracking. (Its a big project... but it really needs to be done. I wonder if I'll ever find the time to do it.) -David Nolan Network Software Designer Computing Services Carnegie Mellon University ___ mon mailing list [EMAIL PROTECTED] http://linux.kernel.org/mailman/listinfo/mon
Alert escalation even if acknowledged
(Sorry if this is a dup. Append resent after 17 hours without confirm.) I'm trying to meet a request to have mon send an alert first to the on call person, and then if the problem is not resolved within a certain time, T2, to a higher level person. The sticky part is that this should occur even if the condition was ack-ed. If not for the ignoring of the ack, I see that I could do this with a second alert entry (under another period label:) using something like "alertafter T2". In order to ignore the ack, it looks like I would have to set up a duplicate group:service with the same test but, again, "alertafter T2" Alternatively, it looks like I could easily tweak the code to add a new alert type, alertnoack, which acts like alert but disregards the ack state (and maybe passes the alert plugin the ack state and comment). I really don't want to go off on my own. Does this option have any merit? An even more elaborate change would have acknowledgment levels and something like an "alertbelowack num" clause that indicates what level of ack squelches that alert (e.g. default is num=1). Are there any other solutions that don't involve configuring a "duplicate" group:service or changing mon? Thanks for any suggestions, Michael Vogt __ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com ___ mon mailing list [EMAIL PROTECTED] http://linux.kernel.org/mailman/listinfo/mon