The console problem: when a platform grows faster than its navigation

There is a stage every growing platform hits where the features are fine and the product is not. Ours arrived in the RMM module: fleet, devices, jobs, scripts, patching, diagnostics — each one built properly, each one a sibling route, and collectively a set of destinations you had to already know about to reach.

That is not a design flaw in any individual screen. It is what happens when capability accumulates faster than the structure meant to hold it, and it is invisible to the people building it, because they know all the routes.

An overview is a claim, not a summary

Consolidating them behind one workspace shell was the easy half. The interesting half was the overview screen, because an overview is where dashboards usually start lying — not maliciously, but through numbers assembled for visual balance rather than derived from something.

The rule we held to: every figure on the overview reports real fleet health and real job state, and if a number is displayed, clicking it lands on exactly the filtered list it counted. A count that does not reconcile with the list behind it is worse than no count, because it teaches people to distrust the screen and then they stop looking at any of it.

A number on a dashboard is a promise that a matching list exists. Break that once and the whole screen becomes decoration.

Deep links are a contract

The other constraint: every previous route had to keep working. Technicians bookmark things, paste them into tickets, and hand them to each other. A restructure that 404s those links has quietly deleted a colleague's notes — and the cost lands on people who had no say in the redesign and no warning it was coming.

Authorization boundaries were preserved verbatim as well. A navigation change is exactly the kind of refactor where a permission check gets moved to a wrapper that turns out to have a different scope, so the shell changed and the checks did not.

What is honestly still rough

This is a bounded first phase. Several child screens inside the workspace are still on their previous layouts, and they are tracked as follow-on work rather than described as finished. A consolidation that claims to be complete while half the screens inside it are untouched is how the next scavenger hunt starts.