---
title: Monitor ephemeral servers for reboots
description: >-
  Monitor ephemeral servers for reboots using the system.uptime metric with the
  diff() function to distinguish new hosts from rebooted hosts.
breadcrumbs: Docs > Monitors > Monitor Guides > Monitor ephemeral servers for reboots
---

# Monitor ephemeral servers for reboots

Ephemeral environments spin up and terminate hosts constantly, which can make it challenging to distinguish new hosts from rebooted hosts.

You can use a metric monitor on the `system.uptime` metric can to address this. The uptime metric is an ever increasing timer which resets to 0 when a host boots up. You can use the `diff()` function with the metric to distinguish between a new server, which has an uptime of 0 (new server), and a rebooted server, which will show a change (diff) from a running uptime value to a 0.

The example shown below captures how you can set this up:

{% image
   source="https://datadog-docs.imgix.net/images/monitors/guide/ephemeral_set_up_2025-05-20.3dea6ec076019ea61e9a8889278a43fb.png?auto=format"
   alt="The New Monitor page in Datadog, showing a metric monitor with a threshold alert set to 0 for the system.uptime metric, averaged by host" /%}

- [Learn how to create a monitor](https://docs.datadoghq.com/monitors/)
- [Configure your monitor notifications](https://docs.datadoghq.com/monitors/notify/)