Server Monitoring

Server monitoring software measures and tracks compute performance by continually collecting data from system resources including CPU, memory, I/O, network, and disk usage. IT operations teams use the data to identify, prevent and resolve performance issues related to resource utilization, downtime and response time and to track changes in server functionality.

Server monitoring solutions should be able to monitor and analyze performance metrics coming from the host, the running processes and the database running on the host. In addition, being able to tie the hosts to the higher level service indicates the impact which the failure of an individual host (or several hosts) has on SLAs, or error budgets.

It’s best to standardize on a metrics framework that provides enough information, without overloading teams with too many metrics. Some common frameworks include Rate Errors Duration (RED) or Utilization Saturation Errors (USE). In general, monitoring and alerting on process or application level anomalies is often beneficial for teams that suffer from tool fatigue, which can be caused by metrics that often change, like CPU.