# The Capacity page

Date a saturation before it happens, and know what that date is worth.

## The steps

1. Open **Capacity**, under **Optimization**.
2. Read **Overall Capacity Health** and **Resources at Risk**.
3. In **Saturation Forecast**, resources are ranked by date, the closest first. Filter by resource type and by horizon.
4. Click a resource: its utilization, its date, its forecast curve and what consumes it.
5. Change tab to focus on one family: **VMs**, **Storage**, **SAN**, **Hosts**, **Clusters**, **Backup Repositories**.

## What a forecast is worth

- The trend is calculated on the history of your collections, **source by source**, at one point per day. A busy afternoon does not make a trend.
- A resource marked **est.** does not have enough history yet: Naoshift applies a default growth to it, and says so. Do not budget on that date.
- On arrays, it takes 5 days of collection before a date appears. Until then the screen says so, never zero.

## Where to start

1. **Critical** resources within 30 days need action this month.
2. For the others, think of the procurement lead time: if a host takes three months to arrive, what saturates in three months is decided today.
3. Before buying, look at [the savings](/en/steer/opportunities-and-recommendations): recovering idle machines often pushes a date back by several months.

## Set the thresholds

**Configure Capacity** sets the critical threshold, the warning threshold, the at-risk window and the default growth. **Export** produces the full list, for a capacity review.

### Everything is marked "est."

Your collections are too recent. With a collector, the trend builds itself in a few days. With files, import at regular intervals.

### The SAN tab is empty

It needs a connected array: see [Add backups and arrays](/en/connect/backups-and-arrays).
