The server ran out of disk space! This disruption along with similar bottlenecks could have been predicted and avoided by using a Windows Performance Monitor.

These monitors actively collect background performance data published by the Windows operating systems and trend the results over time. The slope, shape, and values will indicate threats to performance. In addition, threshold alerts can notify System Admins before the resources reach critically low levels.

The ELM Core License provides three (3) important Windows Performance Monitors.

ELM Dashboard Bottleneck Sparklines

The ELM Dashboard displays five sparklines that identify performance bottlenecks. These sparklines focus on Processor, Memory, Disk, Disk Space and Network metrics. To build these displays, a set of default Performance Collectors are assigned a group of Objects/Counters/Instances. On a scheduled basis, the performance data is collected, weighted and normalized. The most threatening within the group becomes the Bottleneck Value and is displayed for each of the 50 time periods.

When the bottleneck value reaches an 85% of the recommended range, the trend lines are shifted to red. This provides a simple visual representation of the where the hot spots are developing.
Window Performance Monitor in ELM Dashboard displaying Bottleneck Sparklines.
In addition to the sparklines, there is a supporting set of Bottleneck Tables and Charts. The tables list the native performance metrics and the calculated Bottleneck Values. The charts independently trend each of the weight and normalize values.

Windows Performance Monitor

Windows Performance Monitor in ELM Enterprise Manager to alarm on low disk space.

Windows Performance Monitor in ELM Enterprise Manager to alarm on low disk space.

The Performance Monitor is one of the tools included in the ELM Enterprise Manager-Core License. It monitors any performance Object/Counter/Instance value published by the Windows operating systems.

On a specified schedule, it compares the current value against a target value and the selected operator. If the value fails these conditions, ELM writes a Windows event log entry to either the local Application Log or directly into the ELM Primary Database.

Then, utilizing the event log management technologies, filtered Views will display the event and trigger an alert.

Common notification options include elevating the Status Icon in the ELM Dashboard to 7-9 (Red Diamond) or sending a detailed email message. Both methods will be triggered immediately after the event is processed.

Scheduled Performance Reports

To complement the Bottleneck sparklines and threshold monitoring, the reporting feature provides a longer-term perspective. Often run as weekly reports, they document daily averages of key performance metrics.

These reports are scheduled to be run at specific times and include a predefined collection period. They can be delivered by email or filed in a user-specified location. Outputs include .pdf or CSV formats.

The default reports include:

  • Logical Disk IO
  • Logical Disk Storage
  • Memory
  • Processor

The server ran out of disk space…  This is one of the most commonly reported causes of unexpected server downtime. In many cases, they could have been predicted by using the Windows Performance Monitors in ELM Enterprise Manager.