Sorina,
for the custom UI side I let me customers (internal and external) decide what they think is important. As an example I have third party data from smarts and I bundled that into dashboards for network devices for one team, firewalls, and then Hosts (servers of all flavors) into other dashboards. Commonly they will ask for CPU util, memory and disk usage, so I make a big four of heat maps with multiple drop down metrics they want to see in each window for cpu, memory, disk, and maybe network or interface and ports. The storage guys want all the netapp and emc info they can collect and blend that into the vcenter data and you should be able to create many dashboards for the storage teams that helps them with read write rates and latencys are most common.
Dont share it, but I always recommend using the replay dashboard called the metric weather map. You have to have a speicifc metric in mind, or Super Metric but the replay over time is great to see building issues that answer the question of what happened this weekend.
Use your tags and application creation to help make the dashboards.
