Monitor Linux load average

The Linux (or other Unix like systems) load average tells about the busyness of the server. More specifically it is the average number or running processes.

The load average is reported for the last minute, the last 5 minutes and the last 15 minutes.

Monitor Linux load average

You can use DownNotifier to monitor the load average. The code example below reports the load average to DownNotitier every 15 minutes. After putting the script on your server, setting the correct API key and adding it to the cron DownNotifier sends an e-mail about the newly added metric. Use the link in the e-mail or log in to your account to manage the metric and add alerts.

Get an alert with high load average

We recommend setting a “Value is above” alert and the set a longer period, to avoid alerts with a very temporary overload. You can for example choose for:

  • Value is above 1 (multiply it by your number of cores) in the last 15 minutes
  • Value is above 0.5 (multiply it by your number of cores) in the last 8 hours
When is the load average too high?

In general, a lot more than 1 per core means the system is overloaded. So for example a single core system should have a load under 1, an 8 core system should have a load under 8. Use the “nproc” command to see the number of cores.

When the load is too high the bottleneck can be the CPU, but it is also possible the disk or networking is the bottleneck. Use tools like “top” to investigate this.

Get an alert when the server does not report

We recommend setting an (additional) alert to get alerted when the server does not report for a longer period (for example 1 hour). This way you get noticed when there is a problem with the script or the cron daemon. You can use a “Not updated / inactive” alert for this.