Large template fails to apply to host

[jokes54321]

I built a large template to monitor our Juniper Switch stack, which consists of 10 - 48 port switches configured as a single virtual chassis.My template has over 4000 items and 2500 triggers. When I go to apply the template to the host nothing happens.

Is there a limit on how many items could/should be in a template? Any suggestions.

Denny

[heaje]

What do you mean by "nothing happens"? Do you mean that literally nothing changes on your screen when you click apply? Does it go to a screen with only the zabbix menus and header, but no actual content? Does everything look like it ran normally, but the template simply wasn't applied?

Is PHP configured to show errors on-screen? If not, have you checked your web server logs to see if any errors show up there? If PHP is not configured in php.ini to show errors, and no errors are in your web server logs, try setting the following variables in php.ini:

error_reporting  =  E_ALL
display_errors = On

After doing that, try to apply your template again

I've seen this kind of thing in the past due to PHP hitting the max memory limit as it is specified in php.ini.

[jokes54321]

Hello,

I appreciate your response and apologize for being so vague. When clicking save I didn't get any error messages and it did go to a screen with only the zabbix menus and header, but no actual content.

While waiting for a reply I tried taking my one template and broke it into 11 templates, one for each chassis in the virtual stack and one for the LAG interfaces. I was able to apply those to my host and have a new issue. Typically a new issue would warrant a new post, but I am wondering if it's related to the same issue, PHP memory limits.

Between the 11 templates there are 4000+ items I am monitoring. After applying the templates to the host I started to notice the Zabbix queue backing up and the counter for SNMP over 10 minutes was around 7000 and climbing. I assumed this was a resource limit and increased the Zabbix pollers by a factor of 10 along with the MySQL connections setting. This seemed to help a bit, but not much.

While Zabbix was monitoring I decided to use my MIB browser to Walk the interface list and noticed the results were coming back at a snails pace. The moment I stopped the Zabbix service my MIB browser was able to walk the interface list rapidly.

This leads me to believe the problem isn't Zabbix not being able to keep up but the switch stack isn't able to handle all the SNMP polling.

My 4000+ items are all dynamic indexes and I see dynamic indexes are cached as of version 1.6.3. I'm wondering if this cache isn't big enough to cache all of my dynamic indexes causing repeated index look ups.

Is there a setting that controls how big this index cache can be and a way to monitor it?

Denny

[jokes54321]

After posting my long response, I happened to see some of my concerns were addressed in 2.0.5. We're going to apply the update and see how it goes.

Thank you again,

Denny