Regular monitoring checks a site from one point. But a site can work in Germany and not open in Kazakhstan because of providers and local networks. MultigeoMonitoring checks domains from selected countries and sends an alert when something goes down somewhere. The spec came as a wish list; I rewrote it into a BRD and then did the structure, interface and landing page.
MultigeoMonitoring
A service that checks website availability from dozens of countries. I led the product from breaking down the spec to the interface. It now runs as checkme.dev.

Context
Monitoring that looks from different countries
Task
The “domains × countries” matrix
Dozens of countries times hundreds of domains. Show it all at once and you get a heat map where nothing is visible, and at a thousand domains it simply falls apart. It had to be clear within a second where things are bad.
The product opens on problems
Usually the main monitoring screen is a green “all good” dashboard, with errors on a separate tab. I did the opposite: by default you only see what’s down, in which countries and with what error. The on-call engineer doesn’t have to search for anything and immediately sees what to react to.

Country at the core of the model
Competitors hide country selection in advanced settings. Here the country is at the core of the data: each has its own ISO code, flag, time zone and date format. Countries for a domain are picked by region in one window.


Three levels instead of one huge table
I split the heat map into three levels. At the top, only countries with problems. Expand a country and you see the domains that are down. Expand a domain and you see the specific check. That way the screen doesn’t fall apart even at a thousand domains.
The first thing a person sees is what’s broken. Reports and charts come as a second layer.
Positioning
How it differs from UptimeRobot, StatusCake and BetterStack
Outcome
The product launched as checkme.dev
The service is live at checkme.dev. I also designed the landing page, in several versions, but it wasn’t picked up: there were no resources, and the frontend developer built the site himself.
What I’d do differently
I’d test the level-by-level drill-down with real on-call engineers. And I’d put the status colors into tokens with a contrast check right away.