In <UQX708PZ3AFGXC02XC at mailvel.velizy.fr.sterianet> Fr�d�ric Mangeant <frederic.mangeant at steria.com> writes:

some alerts don't work as expected.

[snip config and summary of sent alerts]

I think there are 2 false errors :

• for each 'foo.cpu' alert I got paged twice, with the same ACK code.
• I shouldn't have been paged between 11h05 and 12h05, nor after 13h00, for 'foo.procs'

Could you try running "bbcmd hobbitd_alert --test foo cpu" ?

Also, if you add the option "--cfid" to the hobbitd_alert commandline in hobbitlaunch.cfg, it will include the linenumber of the hobbit-alerts.cfg file with each alert. That should make it easier to track down what rules trigger an alert.

Regards, Henrik

...

I think there are 2 false errors :

• for each 'foo.cpu' alert I got paged twice, with the same ACK code.
• I shouldn't have been paged between 11h05 and 12h05, nor after 13h00, for 'foo.procs'

Could you try running "bbcmd hobbitd_alert --test foo cpu" ?

Of course :

$ $BBHOME/bin/bbcmd hobbitd_alert --test foo cpu 2005-02-15 14:59:22 Using default environment file ../etc/hobbitserver.cfg Matching host:service:page 'foo:cpu:' against rule line 115:Matched *** Match with 'HOST=foo TIME=W:0900:1800' *** Matching host:service:page 'foo:cpu:' against rule line 116:Matched *** Match with 'SCRIPT /tmp/alerte.sh SERVICE=* EXSERVICE=disk,mem,procs' *** Script alert with command '/tmp/alerte.sh' and recipient SERVICE=* Matching host:service:page 'foo:cpu:' against rule line 117:Failed (min. duration) Matching host:service:page 'foo:cpu:' against rule line 118:Failed (color) Matching host:service:page 'foo:cpu:' against rule line 119:Failed (time criteria)

Here are lines 115 to 119 of my $BBHOME/etc/hobbit-alerts.cfg :

115 HOST=foo TIME=W:0900:1800 116 SCRIPT /tmp/alerte.sh SERVICE=* EXSERVICE=disk,mem,procs 117 SCRIPT /tmp/alerte.sh SERVICE=disk DURATION>5m REPEAT=2h 118 SCRIPT /tmp/alerte.sh SERVICE=mem COLOR=yellow REPEAT=24h 119 SCRIPT /tmp/alerte.sh SERVICE=procs TIME=*:1145:1150,*:1205:1300 REPEAT=24h

Also, if you add the option "--cfid" to the hobbitd_alert commandline in hobbitlaunch.cfg, it will include the linenumber of the hobbit-alerts.cfg file with each alert. That should make it easier to track down what rules trigger an alert.

Done.

I just noticed this won't work for SCRIPT recipients, because it's put in the message subject which scripts ignore. So drop that.

Undone ;-)

Regards,

--

Frédéric Mangeant

Hi Henrik

I've tried, but I cannot make this happen on my own setup.

Could you send me the script you use for alerting, and the ~hobbit/data/ack/notifications.log file ?

Well, I moved to another server, on which I cleanly installed Hobbit 4.0-rc2

• patches, and can't seem to reproduice the problem.

Anyway, here's my tiny paging script :

$ cat /tmp/alert.sh #!/bin/sh

DATE=date +%d/%m/%Y%t%H:%M:%S echo "$DATE $BBHOSTNAME.$BBSVCNAME = $BBCOLORLEVEL (ack : $ACKCODE, recovered : $RECOVERED)" >> /tmp/alert.txt

I did some more testing, there seems to be 2 small problems :

1. Warning when the format of a script is missing

With this rule :

$ cat $BBHOME/etc/hobbit-alerts.cfg HOST=fmangeant SERVICE=* EXSERVICE=procs,disk,mem,svcs REPEAT=24h TIME=W:0900:1800 SCRIPT /tmp/alert.sh FORMAT=TEXT HOST=fmangeant SERVICE=disk DURATION>2m SCRIPT /tmp/alert.sh

I get a warning :

$ $BBHOME/bin/bbcmd hobbitd_alert --test fmangeant disk 2005-02-16 15:22:03 Using default environment file /BB/hobbit/server/etc/hobbitserver.cfg 2005-02-16 15:22:03 Ignoring SCRIPT with no recipient at line 2 Matching host:service:page 'fmangeant:disk:' against rule line 1:Failed (service excluded) Matching host:service:page 'fmangeant:disk:' against rule line 2:Failed (min. duration)

If I add the format of the script, like this :

$ cat $BBHOME/etc/hobbit-alerts.cfg HOST=fmangeant SERVICE=* EXSERVICE=procs,disk,mem,svcs REPEAT=24h TIME=W:0900:1800 SCRIPT /tmp/alert.sh FORMAT=TEXT HOST=fmangeant SERVICE=disk DURATION>2m SCRIPT /tmp/alert.sh FORMAT=text

$ $BBHOME/bin/bbcmd hobbitd_alert --test fmangeant disk 2005-02-16 15:22:54 Using default environment file /BB/hobbit/server/etc/hobbitserver.cfg Matching host:service:page 'fmangeant:disk:' against rule line 1:Failed (service excluded) Matching host:service:page 'fmangeant:disk:' against rule line 2:Failed (min. duration)

2. Repeat interval not correctly taken into account

I tried to repeat an alert every 5 minutes :

$ cat $BBHOME/etc/hobbit-alerts.cfg HOST=fmangeant SERVICE=* EXSERVICE=procs,disk,mem,svcs REPEAT=24h TIME=W:0900:1800 SCRIPT /tmp/alert.sh FORMAT=TEXT HOST=fmangeant SERVICE=disk DURATION>2m SCRIPT /tmp/alert.sh FORMAT=TEXT HOST=fmangeant SERVICE=procs REPEAT=5m SCRIPT /tmp/alert.sh FORMAT=TEXT

$ $BBHOME/bin/bbcmd hobbitd_alert --test fmangeant procs 2005-02-16 15:23:59 Using default environment file /BB/hobbit/server/etc/hobbitserver.cfg Matching host:service:page 'fmangeant:procs:' against rule line 1:Failed (service excluded) Matching host:service:page 'fmangeant:procs:' against rule line 2:Failed (min. duration) Matching host:service:page 'fmangeant:procs:' against rule line 3:Matched *** Match with 'HOST=fmangeant SERVICE=procs REPEAT=5m SCRIPT /tmp/alert.sh FORMAT=TEXT' *** Script alert with command '/tmp/alert.sh' and recipient FORMAT=TEXT

But I got paged every 30 minutes :

$ cat /tmp/alert.txt 16/02/2005 14:43:27 fmangeant.procs = red (ack : 145155, recovered : 0) 16/02/2005 15:13:30 fmangeant.procs = red (ack : 145155, recovered : 0)

Is it possible to use any repeat value ?

Thanks in advance.

Regards,

--

Frédéric Mangeant