With IBM Cloud Pak for AIOps v4.12.0, Topology Manager gets a useful scalability bump: we’re now tested at 25 million resources and 60,000 topology groups in a custom-sized deployment. On paper, it looks like “just bigger numbers,” but for anyone running large or distributed environments, it’s actually a meaningful quality‑of‑life improvement. A resource in Topology Manager isn’t just a device; it’s every vertex in the graph: slots, cards, ports, pods, containers, disks, servers, routers—you name it. Modern infrastructures generate a lot of these, and the more detail you can keep, the more accurate your dependency graph becomes.

The real benefit is that this extra headroom reduces the amount of time you spend deciding what to trim or simplify before onboarding topology. Previously, for very large estates, you often had to choose between modelling everything accurately or keeping the system within comfortable limits. That usually meant excluding “less important” layers, flattening parts of the network, or bundling things into higher‑level abstractions—which works, but it also removes context that AIOps depends on for correlation and noise reduction.

With support for 25 million resources and 60,000 groups, you can now include more detail and create more meaningful groups—applications, sites, VPNs, service boundaries—without constantly worrying about scale constraints. The end result is a cleaner, more complete topology that simply behaves better when you need it most.

Also check out the related blog from @Ian Manning at [this link].