Similar to cache usage and process busy rates, we may also monitor the count of unsupported items with an internal item.

To create such an item, follow these steps:

1. Let's go to Configuration | Hosts, click on Items next to A test host, and then click on Create item. Fill in these values:

• • Name: Amount of unsupported items
  • Type: Zabbix internal
  • Key: zabbix[items_unsupported]

2. When done, click on the Add button at the bottom. After a short while, check this item on the Latest data page:

14? That's an extremely high value for such a small installation, although in this case it's caused by the VMware monitoring being down. At this time, a VMware timeout results in all VMware items becoming unsupported. In a perfect environment, there would be no unsupported items, so we could create a trigger to alert us whenever this item receives a value larger than 0. That wouldn't be too useful anywhere but in really small environments, though; usually, a thing becomes broken here or there, and the unsupported item count is never 0. A more useful trigger would hence be one that alerts about a larger increase in the number of unsupported items. The change() trigger function could help here:

{A test host:zabbix[items_unsupported].change()}>5 

Whenever the unsupported item count increases by more than 5 in 30 seconds, which is the default item interval, this trigger will fire. The threshold should be tuned to work best for a particular environment.

Such a global alert will be useful, but in larger environments with more distributed responsibilities, we might want to alert the responsible parties only. One way to do that would be monitoring the unsupported item count per host. With this item, it probably makes most sense to create it in some base template so that it's applied to all of the hosts it's needed on.

Let's create such an item with the following steps:

1. Navigate to Configuration | Templates, click on Items next to C_Template_Linux, and then click on Create item. Fill in these values:

• • Name: Unsupported item count
  • Type: Zabbix internal
  • Key: zabbix[host,,items_unsupported]

2. When done, click on the Add button at the bottom. Check this item on the Latest data page:

Apparently, the test host has 0 unsupported items in this installation. We would now create a trigger on the same template, alerting us whenever a host has a non-zero count of unsupported items. Such a combination would work fairly well although, in larger installations, it could result in a large number of triggers firing if an item got misconfigured in the template or if a broken userparameter script were distributed. Unfortunately, there's no built-in item to determine the unsupported item count per host group. One workaround would be to use aggregate items, as discussed in Chapter 10, Advanced Item Monitoring. For example, to obtain the unsupported item count for a group called Linux servers, the aggregate item key could look like this:

grpsum[Linux servers,"zabbix[host,,items_unsupported]",last] 

We should probably avoid creating a trigger for the unsupported item count on individual hosts, creating one on the aggregate item instead. While the individual items would keep collecting data, which is a bit of a load on the Zabbix server and increases database size, at least the alert count would be reasonable.

There're still more internal items. It's a good idea to look at the full list of available items for the latest version of Zabbix in the online manual.