Hi all.
I'm brand new to Zabbix but I'm an experienced UNIX/Linux admin. I'm trying a small proof-of-concept with 2.0.3 (will go to 2.0.4 soon, if the PoC works out). We're migrating away from HP's SiteScope and I'm evaluating several packages, including Zabbix and Nagios. Although I've read "Zabbix 1.8 Network Monitoring" by Rihards Olups via Safari and I've also been through quite a bit of the Zabbix 2.0 manual and turned up lots of threads in the forums related to templates and dependencies, I still don't understand how to organize/nest templates with dependencies so that there isn't a tremendous amount of duplication.
Here's a scaled down ASCII diagram of our network topology:
I would like to create a few templates for the standard items we monitor and nest/link them so that it's easy to apply a template like "Datacenter R Linux base monitors" to a host and have all the dependencies (triggers) correctly handled by the template(s).
We have historically used a check on whether we can connect to the ssh port to detect whether a host is up (as Linux boxes will respond to ping even when they're otherwise unresponsive), and then any other items for a particular host depend on that.
The problem is that I don't see how I can organize my templates so that I don't end up duplicating large amounts of them for each datacenter. If I want to make the host-specific item "LDAP port" for host "ldap2" in Datacenter R depend on a template for "SSH Port is reachable", it appears I need a separate template for "SSH port is reachable for host in datacenter B", "SSH port is reachable for host in datacenter I", "SSH port is reachable for host in datacenter R". Once I do that, I essentially need to duplicate the trigger for every item for each datacenter.
There has to be a better way, but it's not obvious how I organize this to avoid duplicating many items and triggers for each of the datacenters.
All suggestions appreciated,
Thanks,
Tim
I'm brand new to Zabbix but I'm an experienced UNIX/Linux admin. I'm trying a small proof-of-concept with 2.0.3 (will go to 2.0.4 soon, if the PoC works out). We're migrating away from HP's SiteScope and I'm evaluating several packages, including Zabbix and Nagios. Although I've read "Zabbix 1.8 Network Monitoring" by Rihards Olups via Safari and I've also been through quite a bit of the Zabbix 2.0 manual and turned up lots of threads in the forums related to templates and dependencies, I still don't understand how to organize/nest templates with dependencies so that there isn't a tremendous amount of duplication.
Here's a scaled down ASCII diagram of our network topology:
Code:
Datacenter: B
+-----------------------------------------------------+
| |
| +---------------+ +---------------+ +-----------+ |
| | Zabbix server | | host: ldap1 | | host: ad1 | |
| +-------|-------+ +------|--------+ +-----|-----+ |
| | | | |
| +----------+------+-----------------+ |
| | |
| +-------+--------+ |
| | gateway B | |
| +-------+--------+ |
| | |
+--------------------V--------------------------------+
|
Datacenter: I |
+--------------------V--------------------------------+
| | |
| +-------+--------+ |
| | network core I +------------+ |
| +-------+--------+ | |
| | | |
| +----------+------+ | |
| | | | |
| | | | |
| +-------|-------+ +------|--------+ | |
| | host: dns1 | | host: cifs1 | | |
| +---------------+ +---------------+ | |
| | |
| +--------V--------+ |
| | network core R | |
| +--------V--------+ |
| | |
+------------------------------------------V----------+
|
|
Datacenter: R |
+------------------------------------------V-----------+
| | |
| +-----V--------+ |
| | gateway R | |
| +-----V--------+ |
| | |
| +-----------------+--------------+ |
| | | |
| | | |
| +---------------+ +---------------+ |
| | host: ldap2 | | host: backup1 | |
| +---------------+ +---------------+ |
| |
+------------------------------------------------------+
We have historically used a check on whether we can connect to the ssh port to detect whether a host is up (as Linux boxes will respond to ping even when they're otherwise unresponsive), and then any other items for a particular host depend on that.
The problem is that I don't see how I can organize my templates so that I don't end up duplicating large amounts of them for each datacenter. If I want to make the host-specific item "LDAP port" for host "ldap2" in Datacenter R depend on a template for "SSH Port is reachable", it appears I need a separate template for "SSH port is reachable for host in datacenter B", "SSH port is reachable for host in datacenter I", "SSH port is reachable for host in datacenter R". Once I do that, I essentially need to duplicate the trigger for every item for each datacenter.
There has to be a better way, but it's not obvious how I organize this to avoid duplicating many items and triggers for each of the datacenters.
All suggestions appreciated,
Thanks,
Tim