Technical Service Offering Resiliency Tiers

[zcurtin]

Is it ok to have a hierarchy of Technical Service Offerings?

Context:

We are building out SRE (SLO/SLI) for our Infrastructure Offerings.

We are grouping our Infrastructure Offerings into different 'Resiliency Tiers.' The lowest level of Resiliency for infrastructure is easy: up/down Events.
Once you get past that lowest level it gets much more complicated to point at 1 object which determines whether a given instance of the Technical Service Offering experiences degradation or outage.

For some Technologies it's easy- a set of 3 Servers with a Load Balancer (2 Load Balancers) behind them. We should only care if all 3 of the servers go down OR both of the Load Balancers go down.

For other Technologies it's a lot harder- Mongo Replica sets are a good example- You only care if the entire Replica set goes down.

So we have established that there is not always 1 object you can send a Degradation/Outage event to which means we have to make one in some cases. The only real solution I can find in CSDM is to create a child Technical Service Offering of 1 overarching Technical Service Offering.

This grouping would be created automatically at build time by the Infrastructure build process that's already in place.

So lots of words but the question rolls back to: Is it ok to have a hierarchy of Technical Service Offerings (TSO child of TSO parent) or is there another way to determine Resiliency Offerings?

[Nachiketh Rahul]

I guess the answer is yes and no depending on how you want to structure this. You could surely use a parent child structure but it still does not give you visibility into RPO/RTO etc. - Ideally the resiliency and recovery is tracked under business continuity in servicenow - 

Below link might help - 

https://docs.servicenow.com/en-US/bundle/vancouver-governance-risk-compliance/page/product/grc-busin... 

[zcurtin · Accepted solution]

For anyone wondering about my question I have an answer from Mark Bodman.  The correct class to represent these groupings is Application Service.  This being the right class really underscores the need to switch to 'Application System' as the name of the class (Mark is the one making this point).