Server Health Monitoring
Server Health Monitoring is fundamentally concerned with observing how a server (be it bare-metal or a virtualized instance) reacts to the operating load placed upon it. Wormly Metrics makes this easy with collectd integration for Linux, BSD et al, Windows, and also at the application level.
On Linux it's as easy as installing collectd like this:
curl https://www.wormly.com/collectd/API_KEY/HOST_ID | bash
Such monitoring assists with both capacity-planning; preventing server failures by ensuring instances retain sufficient capacity to conduct the required tasks including peak loads.
It also assists with anomaly detection - whereby unusual or unexpected behaviour is noticed, and can be mitigated before it results in failure.
Server Health Monitoring provides a very useful additional perspective when combined with Uptime Monitoring; it assists you in preventing downtime with informed capacity planning, rather than merely being able to react to failure events.
Common scenarios resulting in a server failure include excessive CPU usage, insufficient RAM and excessive Disk IO operations. By monitoring these metrics - and others - you can send alerts when health metrics exceed thresholds you define. For example:
On Linux, BSD, and UNIX-like platforms, we integrate with collectd to push useful health and performance data to the Wormly Metrics service.
For Microsoft Windows servers & instances, health monitoring is facilitated by the open source BBWin agent.
In both instances, this technique is known as agent-based monitoring, which allows an external service such as Wormly to monitor internal server health and performance in a secure manner.
The data collected is summarized and presented in graph format enabling you to easily observe the performance trends of many critical system parameters.
System Performance Metrics
Learn more about the performance metrics monitored and graphed by Wormly:
Also available “agent-less”
You may also elect to use Wormly without installing the agent on your server. In this case you still have access to all externally measurable performance metrics and their associated trend graphs.