Hi there!
Templates:
I would be very interested what you guys use? Do you use most of standard-templates or are you building all templates new on your own?
We design every template new; do not use any of the given standard-templates. And it´s really really hard work.
If I had to sum up, I would guess:
- 2% installation and fine-tuning zabbix-server / zabbix-proxies
- 20% getting and fine-tuning templates for basic infrastructure zabbix-server and -proxies with all dependencies betwenn them (ups-systems, switches, routers, wan-connections, etc.)
- 75% writing our templates for all kind of stuff (hardware, network, apps, the whole thing we admins have to deal with every day,...)
- 3% get new coffee or something,...
The "problem" we see is the following: you can only monitor something, if you REALLY understand, how it works (find the right items and trigger-expressions) and what / how it acts / cooperates / depends with other things (trigger-dependencies).
Saying that, the logical next step is: all your knowledge is at the end of the day somehow integrated in your monitoring-system.
What do you think about that? How is it handled in your company?
Are you aiming a "perfect" monitoring on your site, that tells you quite exactle whats going on or better NOT going on (and only the one thing -> trigger-dep. / nodata :-))?
Or are you happy with ping- / service-down-messages?
Kind regards,
Markus.
Templates:
I would be very interested what you guys use? Do you use most of standard-templates or are you building all templates new on your own?
We design every template new; do not use any of the given standard-templates. And it´s really really hard work.
If I had to sum up, I would guess:
- 2% installation and fine-tuning zabbix-server / zabbix-proxies
- 20% getting and fine-tuning templates for basic infrastructure zabbix-server and -proxies with all dependencies betwenn them (ups-systems, switches, routers, wan-connections, etc.)
- 75% writing our templates for all kind of stuff (hardware, network, apps, the whole thing we admins have to deal with every day,...)
- 3% get new coffee or something,...
The "problem" we see is the following: you can only monitor something, if you REALLY understand, how it works (find the right items and trigger-expressions) and what / how it acts / cooperates / depends with other things (trigger-dependencies).
Saying that, the logical next step is: all your knowledge is at the end of the day somehow integrated in your monitoring-system.
What do you think about that? How is it handled in your company?
Are you aiming a "perfect" monitoring on your site, that tells you quite exactle whats going on or better NOT going on (and only the one thing -> trigger-dep. / nodata :-))?
Or are you happy with ping- / service-down-messages?
Kind regards,
Markus.
Comment