When to trigger an alert when monitoring Matomo servers?
As use of your Matomo instance increases over time, so too will the utilization of system
resources, like CPU, memory, and storage. It’s important to configure monitoring and alerting so that you’re aware of potential issues before they become critical enough to negatively impact application performance or availability.
With a monitoring solution of your choice (such as Zabbix, Munin, Nagios), it is recommended to monitor the following resources:
● General system health such as CPU and memory usage.
● Storage usage, throughput, and latency.
● Network throughput, number of clients, and error rates.
● Database usage and throughput.
It is recommended the following alerting thresholds as a starting point of what are the most
important components to monitor. These alerts should be created on each server in the cluster.
Type | Severity | Threshold |
---|---|---|
Disk storage | Warning | Disk use exceeds 70% of total available. |
Disk storage | Critical | Disk use exceeds 90% of total available. |
CPU and load average | Warning | 15 minute load average exceeds 1x CPU cores. |
CPU and load average | Critical | 15 minute load average exceeds 2x CPU cores. |
RAM Memory | Warning | Sustained RAM usage exceeds 50% of total available memory |
RAM Memory | Critical | Sustained RAM usage exceeds 70% of total available memory. |