Capacity Dashboard

A capacity planning dashboard for infrastructure teams, built by Emre Erokyar. See how much compute, memory, storage and network you use, what happens when a host fails or demand grows, and when you will run out of headroom.

Open the dashboardContact Emre

What it does

• Compute, memory, storage and network utilization across regions, environments and applications, with history up to five years
• Host failure, demand growth and new-application scenarios, with recovery options: move a workload or add capacity
• Forecast and what-if planning against an 85% planning threshold, with cost estimates
• An Executive briefing for decision makers and an Operations view for day-to-day work
• English and Czech, light and dark themes, keyboard accessible, works on phones

Frequently asked questions

What is Capacity Dashboard?

A capacity planning dashboard for infrastructure teams. It shows how much compute, memory, storage and network you use across regions, environments and applications, what happens when a host fails or demand grows, and when you will run out of headroom.

Who is it for?

Platform and infrastructure leads and operations teams who watch utilization day to day, and IT and finance leaders who need a plain-language briefing (the Executive view).

What is the difference between the Executive and Operations views?

Operations is the working view: every chart, table and scenario, with sections you can rearrange using Edit layout. Executive is a fixed one-page briefing covering risk, business exposure, cost and the next decision.

What does “at risk” mean?

A system is at risk when any resource (compute, memory or storage) is at or above 85% of its capacity. A host failure removes 25% of the affected system’s capacity, which is why systems close to the threshold show the biggest impact.

What is a storage pool?

A storage pool is the storage that belongs to one system. Opening a pool shows that pool’s storage only; compute and memory belong to the system itself, which you can open from the same window.

Where does the data come from?

The estate you see is a simulated environment of 40 systems and 137 virtual machines across three regions, so every feature can be explored without connecting any infrastructure. All data access sits in one module, so connecting real sources changes the data layer, not the interface.

Can it connect to my own infrastructure?

Not yet. Connecting real sources is the next step: read-only metrics and inventory from cloud providers, monitoring exports and a CMDB. If you have a specific source in mind, get in touch.

How are forecasts calculated?

Forecasts apply a monthly growth rate to current utilization. You can adjust the growth, add capacity and see when the 85% planning threshold is reached. The shaded band widens over time to show uncertainty; it is a planning range, not a statistical confidence interval.

What is it built with?

React and Vite for the interface, Recharts for charts, and a Cloudflare Worker behind the contact form. It is served from Cloudflare’s network and tested with Vitest.

How is quality checked?

Dozens of automated tests run in a few seconds: data rules, every view, tables, charts, pop-ups, accessibility basics and design rules for typography, spacing and color. Every change has to pass them before it ships.

Is it accessible?

It is designed to be. Everything works with the keyboard, pop-ups keep focus inside and return it, charts have text alternatives and zoom buttons (not only dragging), motion respects the reduced-motion setting, and colors keep a consistent meaning. A formal audit is still to come.

Does it work on phones?

Yes. The layout adapts and tables scroll sideways. The dense views are best on a larger screen.

Will it handle thousands of systems?

Today everything is calculated in the browser for 40 systems. At a much larger scale, aggregation would move to a server with paging and virtualized tables; the interface is structured so that can be added without a redesign.