We'll create a separate host, which will be the starting point for the discovery. This host won't do anything else for us besides monitor generic VMware parameters and discover all other entities.

Follow these steps to setup our monitoring in Zabbix:

1. Go to Configuration | Hosts and click on Create host.
2. Enter VMware in the Host name field, clear out existing groups in the In groups block, and enter VMware in the New group field.
3. Switch to the Macros tab and fill in values for these three macros:

• • {$URL}: The VMware API/SDK URL in the form https://server/sdk
  • {$USERNAME}: The VMware account username
  • {$PASSWORD}: The VMware account password

4. Switch to the Templates tab, start typing vmware, choose Template VM VMware, and click on the Add control in the Link new templates section.
5. When done, click on the Add button at the bottom.

What's next? Well, nothing. If everything has been done right, everything should be monitored automatically. Hypervisors should be discovered and monitored, and virtual machines discovered, placed in groups based on hypervisors, and monitored as well. It might not happen immediately, though. Like other Low-Level Discovery (LLD) rules in default templates, VMware discovery also has a one hour interval—wait, LLD rules? Yes, VMware discovery also uses LLD functionality. We discussed it in detail in Chapter 11, Automating Configuration. VMware support takes it a step further, though: besides item, trigger, and graph prototypes, it also uses host prototypes. We'll cover host prototypes a bit later. For now, we can either leave the discovery to happen, or we can go to Configuration | Templates, click on Discovery next to Template VM VMware, and reduce the update interval for all three discovery rules. Just make sure to set it back later.

After waiting for a while—or after reducing the intervals—check Configuration | Host groups. You should see several new host groups, prefixed with Discover VMware VMs. Depending on how large the monitored VMware instance is, the new group count could be from two up to many. There'll be a group called Hypervisors and a group for virtual machines per cluster. If there're clusters, there'll also be a group for hypervisors per cluster.