Capacity
What you have, what you are using, and when you will run out.
FinOps → Capacity is about headroom: what exists, how much is used, and when the trend meets the ceiling.
What is tracked
| Resource | From |
|---|---|
| CPU and memory | Agents |
| Disk, with time-to-full | Agents |
| Database connections | MySQL, ProxySQL |
| Warehouse storage | Connections |
Time-to-full beats utilisation
A disk at 71% is not interesting. A disk at 71% that was at 40% this morning is.
Prediction is what gives you the one thing a percentage threshold cannot: warning while there is still time to do something other than delete files at speed.
Alerting on the prediction is a fleet rule.
Connection headroom
Database connection limits produce one of the most confusing outages there is: everything is healthy, nothing is overloaded, and the application cannot connect. Tracking headroom against the configured limit is how you see it coming.
Planning
The useful question is not "what are we using" but "when does the trend meet the ceiling". Capacity projects the current growth against the current limit and gives a date.
A date is actionable in a way a percentage is not — it can be put next to a procurement lead time.
Related
- Cloud spend — capacity you are paying for and not using
- Servers — per-machine detail
Where this behaviour lives: backend/internal/worker/capacity_worker.go, frontend/src/app/capacity/. If the code and this page disagree, the code is right — please fix the page.
Part of Analytics and cost — Dashboards, SLOs, reports and what the infrastructure costs.