Guidelines for scoping an enterprise-wide ITOM Discovery and Service Mapping
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
2 hours ago
Hi all,
I’m working on an ITOM implementation (Discovery + Service Modeling) for a large institution, and I’d like to hear how others approach this kind of engagement at company level, rather than application by application.
Where we are
We’ve started modeling services, and the question that keeps coming up is how far to go in identifying what sits behind each service. In one case, the suggestion was to list every URL/endpoint a portal calls so we can uncover the underlying structure. That raised an obvious concern: how do you keep that maintained when endpoints change constantly? It made us realize we probably need a clearer, company-wide approach before going deeper.
What I’d love guidance on
- Where to start. For a full-company rollout, what’s your recommended sequence? Infrastructure discovery first, then application services, then business services? Or start with a few critical business services and work down?
- What to ask for, and from whom. When engaging application, platform and integration teams, which inputs do you request up front? (Entry points, API gateway definitions, load balancer/ingress configs, architecture diagrams, existing inventories, APM access...) What do you consider essential versus nice to have?
- Entry points. How do you decide what counts as an entry point for Service Mapping? Per URL, per API, per backend, per application? How do you avoid an unmanageable number of them?
- Depth and granularity. How do you decide where to stop modeling (for example, workloads vs pods, integration servers vs individual flows)? Is there a rule of thumb that has worked for you?
- Completeness criteria. What does “complete” mean in practice for a service map? How do you agree this with stakeholders so it doesn’t turn into “every endpoint must be listed”?
- Maintainability. What mechanisms keep the model alive over time (scheduled refreshes, CMDB Health, data certification, integration with CI/CD, Change Management triggers)? And how do you avoid burdening teams with notifications for every minor change?
- Complementary sources. When do you rely on Discovery and Service Mapping versus Service Graph Connectors, APM or declarative sources such as repos and API definitions?
- Pitfalls. What are the most common mistakes in this type of implementation, and what would you do differently?
Any frameworks, checklists, onboarding templates or lessons learned would be very welcome. Thanks in advance!