Broadcom Software Academy Blog

Fantastic Universes and How to Use Them

Written by Guenter Di Grossberger | Oct 4, 2024 3:38:43 PM
Key Takeaways
  • Learn what universes are, and how they can help you work with DX APM.
  • Leverage DX APM to assign metrics and topology components to one or more universes.
  • Use regular expressions to assign agents to a universe.

You may have seen a menu item or heard one of my colleagues talk about universes in DX APM or DX Operational Intelligence. But what are universes and how can they help you work with DX APM?

In short: Universes filter APM data and implement role-based access control (RBAC). Compared to domains, they allow you to assign metrics and topology components to one or more universes.

Classic CA APM: Domains

If you have been working with classic CA APM 8, 9, or 10, you might be familiar with the concept of domains. Domains were used to split all the data that agents report into CA APM into distinct sets. A domain could stand for an application, a business service, a group or division within the company, or a technology.

A day in the life: Alex and CA APM domains

Alex is responsible for IBM MQ. His job is to make sure that all servers and queue managers are up and running and don’t have performance issues, such as backed up queues. The whole MQ environment at his company is monitored by DX APM. But he does not care about all the other stuff that is also monitored by APM. He wants to focus solely on his MQ environment.

In classic CA APM, Alex (or an APM admin) can define a domain using regular expressions that match only the agents monitoring IBM MQ. As hostnames are arbitrarily assigned in his company, Alex makes sure that the “process name” of all the APM agents monitoring MQ start with “mq”. He has established a naming convention.

Alex then defines the MQ domain with <agent mapping=”.*\|mq.*\|.*”/>. The APM agent name consists of three parts separated by a “|” character:

  1. The hostname, usually read directly from the operating system
  2. The process name
  3. The agent name

These are also the first three levels that you open in the metric tree. The process name and agent name can be defined in properties of the IntroscopeAgent.profile or via environment variables. They can and should be used to implement a naming convention.

In classic CA APM, users and groups can be assigned to a domain with read or write roles. Only people assigned to a domain can see the agents, their metrics, and traces. Domains are mutually exclusive and first come, first serve: An agent is assigned to the first domain in the domains.xml file, where its agent name triplet matches the regex. If there is no catch-all domain at the end, an agent might not be visible at all.

DX APM: Universes

Universes in DX APM are the successor to domains. In the definition of a universe (in Settings/Universes), you can also use regular expressions to assign agents (“metric sources”) to a universe. You can also manually select agents from the list of all agents. I would strongly recommend to use regex because they are generic and work for every new agent as long as it matches the regex.

Unlike domains, universes can overlap. An agent belongs to every universe in which its agent name triplet matches the regex. If you want to give Alex access to two universes, you can either add his user account (or group) to the “Access” section of both universes or you can create a third universe that combines the regular expressions from both universes. In DX APM, you can assign AD or SAML groups to universes. Or, you can allow access to all APM data and use universes to filter to the task at hand.

Universes in DX APM unlock even more power! For example, they allow you to assign distinct management modules to a universe and you can filter the (topology) map components shown in a universe. A topic for another day!