Are Dynamic CI Relationships Possible OR Is It Not Needed?

[Robert Campbell]

Questions:

Is it needed? 

Maybe we're thinking about things wrong and we and need a new perspective.

Use Case

• We want to see infrastructure that supports this service AND infrastructure used to support the infrastructure that supports the infrastructure.
  • Normally we would relate a database instance but due to a lack of understanding of the CMDB and a visual that shows the overall impact, we also relate the server that supports the database to the service because when trying to view more than one-level, things get confusing to some.
• Service to Service (S2S) relationships.
  • Similar to the infrastructure issue, when a service is dependent on a "middleware" service, we relate the middleware and the second-level service because not only if that middleware (API Platform, MQ, Data Warehouse) goes down, you are impacted but if the data you rely on from the other end of that "middleware" goes down, you are also impacted. So everything is put as a first-level relationship.

How to do it?

We would love to be able to automate these relationships but some find it extremely hard to accomplish this task. My suggestion was to create tags or some value in a config file that would identify the parent(s). They didn't like that idea and I haven't been able to come up with another solution. Also, I don't know that we can account for all classes. While moving to the cloud helps with some of this, there is always the S2S relationship that always seems to require manual processing.

Is this even necessary? Should we only focus on service maps? 

My thought is we definitely need to build out the dependencies but others are arguing we only need to use the service map since it's more reliable. 

As I'm writing I'm thinking... Can we use the service map to automate the relationships? As the service map updates, it (a script of some sort) adds and removes the relationships from the relationship table?