# Metrics and activity
Source: https://docs.infraista.com/en/compute/instances/metrics-and-activity



## Metrics [#metrics]

The **Metrics** tab shows graphs of:

* **CPU**: percent of all cores;
* **Memory**: percent in use, with gigabytes used;
* **Network**: traffic in and out on the public interface, in Mbps;
* **Disk I/O**: reads and writes in MB/s;
* **GPU** utilization and VRAM, on AI servers.

Each graph shows the average and the peak for the time range you pick. Usage is sampled every 5 minutes
and kept for 30 days.

<Shot name="compute/instances/metrics" alt="The Metrics tab with CPU, memory, network and disk graphs" />

Right after a server is created, the tab says **Collecting usage**: the first numbers appear shortly
and the graphs fill in over the next hours.

Want to know right away when a server needs attention? Turn on **Instance updates** in
[Notifications](/en/account/notifications).

## Activity [#activity]

The **Activity** tab lists everything that happened to the server, newest first: power changes,
network and firewall changes, billing changes, password resets, snapshots and resizes. Each entry says
who did it: a teammate, **Infraista** (automatic actions) or **Support**.

<Shot name="compute/instances/activity" alt="The Activity tab listing events such as the server being created and firewall rules being changed" />

Filter by kind (**Power**, **Network**, **Billing**, **Access**, **Snapshot**, **Lifecycle**) to find
something quickly. Events are kept for 90 days. Sign-ins, API keys and team changes are in the
organization's [Activity log](/en/account/activity-log).

## Related [#related]

* [Your server's page](/en/compute/instances/your-server-page)