need an example of something simple

[rosen]

okay, i guess there's no configuration file. i guess all the configuration
is done through the web interface.

i'm getting pretty cranky about the manual. i have chapter 4 in front of
me but i have no idea what to do w/ it. chap 4 starts by telling me the
development environment. that doesn't help me at all. next it tells me
what an action is, in terms of ten other things it hasn't defined. and so on.

let's say all i want to do is find out when machine xyz's /VOLUMES/splarg
file system is over 70%.

i can't tell where to start. do i need items? triggers? because it looks like
i can make a trigger and/or an item but how do i connect them?

i need something simple. i see these online "examples" that are 16-20 steps long but #4 makes references to a pulldown i don't have.

i'm a newbie.

[zabbix_zen]

"#4 makes references to a pulldown i don't have"

I've no idea what you mean there, but yes, you do need Items,

think of Zabbix flows in the following way:

Item -> Trigger -> Action -> User/User Group

You have to correctly setup each one of them in that order.

When you reach the User setup don't forget to give read permissions(to the User Group the User belongs to) over the Item/Item Group you want to receive alarms of!

[MrKen]

Perhaps the best place to start is by monitoring your own Zabbix server.

Install the zabbix_agent, it might link to the Template_Linux by default, but if it doesn't, create the host and link it to the Template_Linux. Enable the host.

You can then see in Configuration of Items, the items that are associated with that template (eg Disk space, cpu, running processes, etc).
There may already be some triggers attached to that template, take a look at them to see how they are set up and how they relate to the items.

You can see the data in Monitoring --> Latest data.

Just to expand on what zabbix_zen wrote: The Zabbix heirachy is like this:

Host: Your device, server, whatever, to be monitored.
Item: The parameters or metrics you want to monitor.
Trigger: The Tripwire. The line not to cross. i.e diskspace >70%
Action: What to do if a trigger fires. Send alert? Execute script? etc.

You can also find some better examples on the wiki http://www.zabbix.com/wiki/

HTH
MrKen

[rosen]

here's what i tried.

ZABBIX 1.6.6
for me, jgs == ZABBIX Server.
based on this (alone):
jgs# pgrep -lf zab
12517 -tcsh /opt/depot/zabbix-1.6.6/sbin/zabbix_server
12514 -tcsh /opt/depot/zabbix-1.6.6/sbin/zabbix_agentd
i assume i have the server and agentd running.

i'll give you the output of the web page. however, some of the lines
are long and i'm going to try to keep things under 80 chars. so they'll
look funky.

the wiki says to make a host.

configuration --> hosts

this says:
name dns ip port templates status avail. err acts
ZABBIX jgs 10.3.16.32 unknown
10050 solaris monitored - select

ZABBIX is short for ZABBIX Server.

okay, i guess.

this machine is fine right now.
so if i monitor (almost anything) in the solaris template i
won't get any output, because there's nothing wrong with it.
okay, so i'll make my own item. i'll send out a warning any
time the free space in / drops below 95%. that'll guarantee
me some output, since / is at 35% or so now.

so let's try to make an item.

configuration --> items
descr key update hist trend type status apps error
anything 30 90 265 zabbix agent(active)
vfs.fs.size[fs <,mode>] active - -

configuration --> triggers

from previous thrashing, i have a trigger
anything {ZABBIX Server:vfs.fs.size[/,pfree].last(0)}<95

event generation is normal
severity warning

the trigger page says:


Error:
Trigger just added. No status update so far.

i dunno what that portends.

configuration --> actions
name well
source triggers
escalations not checked
using default subject and msg.
no recovery msg
enabled.

/ is 35% used. so i ought to get something, but i'm not.

i know you're trying to help, and thanks, but i'm getting frustrated.

[MrKen]

Error:
Trigger just added. No status update so far.

i dunno what that portends.

It just means that the trigger has just been added and that the status has not changed. Normally this is no problem, but in your case it would seem that the trigger is not working.

configuration --> actions
name well
source triggers
escalations not checked
using default subject and msg.
no recovery msg
enabled.

/ is 35% used. so i ought to get something, but i'm not.

i know you're trying to help, and thanks, but i'm getting frustrated.

It's not frustrating, it's challenging .

This is the trigger that I use, and I know it works:
{Server_Zabbix:vfs.fs.size[/,pfree].last(0)}<10

What if you were to add a condition to the Action like "Trigger value = PROBLEM".

The only other thing I noticed is that your item is using 'zabbix agent (active)'. Have you tried 'zabbix agent' (no active) just like on the Solaris and Linux Templates?

MrKen

[rosen]

this may be a better hint at my problem(s)

added another machine. jgstoy.
the agentd is running on jgstoy.
it's a linux box.

configurations --> host
dns, ip, template = linux, monitored

configureations --> items

okay, here's my first question. the items list has the one
thing i added, plus all the stuff from the solaris template.
it has nothing from the linux template. the status column
says everything is active.

that seems wrong. it's utterly bizarre to me that i can't
seem to associate the items list with a particular host.
or, how do i do that?

j.

[rosen]

a look at my dashboard.

to amplify on my earlier post:

jgs# ps -ef | grep zab
zabbix 3196 1 49 15:38:59 pts/5 3:00 -tcsh /opt/depot/zabbix-1.6.6/sbin/zabbix_server
zabbix 3200 1 48 15:39:16 pts/5 2:42 -tcsh /opt/depot/zabbix-1.6.6/sbin/zabbix_agentd

so, the server is running, right? but monitoring--> dashboard says

zabbix server .... no (under the running column)
so how do i get this thing to play nice?

thanks again. you guys are trying to help me. i'm frustrated.

j.

[rosen]

and my logs file is empty.

my zabbix log file is empty. i have log level set to 4.
i've restarted the server. i can truss the server and see that it's doing
something.

i have no idea what i'm doing wrong.
j.

[rosen]

i finally figured it out.

the web interface was telling the truth, the zabbix server wasn't running.
(sigh.)

and neither were the agentd's, for that matter. i had bugs in all of
my startup scripts, such that what ps -ef | grep zabbix was showing
me was the startup script trying to start whatever, and not the whatever
actually running.

mea culpa maxima.

thanks for helping. i was blaming zabbix because of a typo in my unix
script that got copied everywhere.

now i'm starting to get the hang of zabbix. i even know why my linux machine
seemed to have the solaris template. problem behind the keyboard.
sorry for the waling.