Custom check monitor

Overview

Custom monitors encompass any service checks that are not reported by one of the out-of-the-box integrations included with the Agent.

Refer to the Agent checks guide for detailed information on writing your own checks that send metrics, events, or service checks.

Configuration

1. Select your custom check:
   

2. Select host or tags that you would like to monitor.
   

   The check runs for every unique set of tags from all monitored hosts. For example, the Nginx service check reports one status per {host,port}. So if you have multiple servers running on a single host, then each one alerts separately in the case of failure.

3. Select your alert options:
   

   While each check run sends a status of either CRITICAL, WARNING or OK, you can choose at what consecutive conditions to cause a state change and a notification. For example, you might want to know immediately when your check fails and only have it recover if it stays that way. In this case you might choose to notify on 1 critical status, 1 warning status and 4 OK statuses.
   You can optionally notify on no data after a configurable timeframe. You must choose at least 2 minutes for your timeframe.

4. Configure your notification options:
   Refer to the Notifications dedicated documentation page for a detailed walkthrough of the common notification options.

Further Reading