It sounds like hierarchical alerting is not active or this is the open source version of Hyperic where this feature does not exist.
The purpose of hierarchical alerting is to avoid triggering alerts for every resource affected by a single root cause.
This setting controls whether alerts are evaluated using the hierarchical alerting method. When hierarchical alerting is enabled, before an alert is triggered for a resource, vCenter Hyperic considers the availability and alert status of the resource's parent.
You can extend the effect of hierarchical alerting by configuring the relationship between a network device or virtual host and the platforms that depend on it, using the Network and Host Dependency Manager.
We have the appropriate version of Hyperic. It's version 5.8.2 vApp downloaded from the vCloud Suite download on the VMware site.. I can see that Hierarchical Alerting is enabled in HQ Server Settings but I can't do anything in Network and Host Dependency Manager. According to the document you sent I should be able to Add / Remove dependents there but there isn't much there to work with. I just see a tab for Top Level Platform and Dependent Platform. I enter a name of a platform in Hyperic in the name field but it never brings anything up.
I opened a case with VMware. The problem is that I'm trying to manage all alerting from vCops. vCops just pulls data from Hyperic without any intelligence. Hyperic has the hierarchical alerting which functions if you setup the alerting to come from Hyperic. I was setting up the alerting in vCops to monitor the availability of services and servers in Hyperic. It's just pulling the information from Hyperic without any that hierarchical alerting functionalilty so it spews out email on services being down in addition to the server being down. Kind of sucks as I wanted to manage all alerting from vCops because it has scheduled maintenance filters while Hyperic does not. These products lack a lot of functionality for the price that you have to pay for them.
vROps 6.0 tackles the situation in a different manner. While it does not have a hierarchy of alerts it introduces a multi symptom alerting mechanism. So instead of having an alert for each service the best course of action would to create a single alert on the server with symptoms from the relevant services. While this approach may require more planning when creating the alerts, it does provide you with the power to group, structure and add meaningful recommendations.