Fifty instances across twelve countries. Each one monitoring several environments. vCenters, NSX deployments, AWS accounts, Azure subscriptions, physical devices, and more. Each environment running hundreds of nodes and hosts. Each generating alerts, warnings, critical problems. Data flows across all of it numbering in the thousands. Open problems reaching into the hundreds of thousands.

*Everything is doing a lot of heavy lifting. You’re trying to reconstruct a globe from fragments, one tab at a time. You log into the next instance. Just business as usual.

VMware’s vRealize Network Insight was an enterprise tool for monitoring and troubleshooting virtual network infrastructure. Within a single instance, it worked well. But enterprises don’t run single instances. They run dozens or hundreds, and vRNI had no way to connect them.

In 2021, VMware decided to build Federation: a system that links isolated instances and surfaces a combined view of an enterprise’s entire network.

Even with federation as a concept, how do you actually show it to someone?

A network admin opens this dashboard and sees information from fifty instances across twelve countries. Where are they? What are they looking at? What do they do first?

This was a wayfinding problem.

The infrastructure was enormous, with data structured differently across instances. Without clear orientation cues, a unified dashboard is just fifty tabs collapsed into one screen.

The data feeding this dashboard is messy by nature. Environments have different data structures, licensing tiers change what’s visible, and feature flags change which widgets appear. Some visualizations are too expensive to compute at this scale.

I personally considered accessibility a non-negotiable from the start.

The design organizes around three questions, in sequence.

The severity colors in the matrix are deliberately different from the dot indicators used elsewhere.

The standard design system colors are too attention-grabbing for a large matrix. They would make the whole panel feel like an emergency.

Instead, the matrix palette was tuned so each severity level is distinguishable at a glance while keeping the whole thing easy to scan.

The alert badges on flows are yellow, not red.

In a federation with thousands of flows, at least a few will always need attention. If those badges were red, the dashboard would permanently look like something is broken.

The color red, and the word ‘problem’ is reserved for when something genuinely is one.

Both the environments and deployed instances modules offer alternate views. Environments can switch between matrix and list view. Deployed instances can switch between map, matrix, and list view.

This serves two purposes: different admins have different mental models for parsing the same data, and alternate views are part of the accessibility layer. Screen readers, keyboard navigation, and WCAG 2.1 compliance were built into every view.

I also designed a modular tooltip system: a single pattern and interaction that serves different purposes in each module.

In the summary, it shows alert breakdowns. In the matrix, device counts. In the map, instance details. Same structure, different content based on context.