Capacity
The Capacity view answers a single question: “based on the last N days of history, when will I run out?” It applies linear regression to your server metrics and turns the slope into a days-until-threshold number you can plan against.
Open it from the sidebar under Monitoring → Analytics → Capacity, or at /capacity.
What you see
Section titled “What you see”Header KPIs:
- Warnings — forecasts trending toward a threshold within an uncomfortable window.
- Critical — forecasts about to hit the threshold shortly.
The main list shows every active forecast, ordered by days-until-threshold ascending:
- Server — the host being forecast.
- Metric — disk (per mount), memory, connections, and other capped resources.
- Current usage — where we are today.
- Trend — daily growth rate.
- Days until threshold — how long until the metric reaches the alert threshold.
- Severity — stable, warning, or critical.
Filter by server, metric, or severity via the toolbar.
What you can do
Section titled “What you can do”Capacity forecasts are read-only projections. From a row you can:
- Open the server to see the underlying metric and act — add disk, tune the app, or archive data.
- Route to alerts — the Alert Rules engine can watch capacity forecasts directly and notify before the threshold is hit.
How it works
Section titled “How it works”A background job crunches the last few weeks of history for every server metric that has a natural cap (disk usage %, memory %, connection budget, etc.), fits a straight line, and projects when that line will cross the alert threshold.
- No trend, or a decreasing trend → stable (no exhaustion date estimated).
- Slow growth → warning, with a large days-until number.
- Fast growth → critical, with a small or single-digit days-until number.
The forecast recomputes on a schedule, so a change in behaviour — a cleanup job, a data migration, a growth spike — shows up on the next run.