SAGE News

Third in a three-part series, guest blogger and ;login: author, Matthew Sacks, presents his review of Open Source Enterprise Monitoring Tools. Matthew looked at Zenoss, GroundWork, and Hyperic.

Next up, Hyperic.

Why we tested these:
There are a number of open source monitoring tools out there boasting large feature-sets and powerful management utilities that go beyond the traditional MRTG, Cacti, or Nagios installation. These monitoring tools utilize these classic open source utilities and build on them to cultivate a full-features enterprise-class monitoring solution. Although these are not the only monitoring tools out there, they are the key players in the open source enterprise monitoring arena.

What was rated:
– Installation process
– Ease of configuration
– Number of features
– Documentation

And now on to the review: Hyperic

Installation:
Hyperic HQ’s installation is straightforward, with the execution of a few simple shell scripts and prompts for configuration data, and the Hyperic server is up and running. Hyperic did well on this installation process by making it very smooth and streamlined. Hyperic requires the installation of an agent on hosts to be monitored. This is an extra configuration step and overhead on the machine, although very minimal, but it allows for seamless collection of information on monitored servers and hosts.

Configuration:
Hyperic is configured by installing an agent on the target host; and then running an auto-discovery from within Hyperic HQ. Aside from the agent installation, most of the configuration within Hyperic is organizing servers into groups and categorizing them. Hyperic can be configured with Nagios as well. Hyperic also has an alerting scheme where events can notify via email, SMS, or SNMP trap. Setting up Hyperic and configuring the agents is mostly a breeze, unless the user requires more fine-grained features. This is the beauty of having an agent-based monitoring tool.

Features:
The key word with Hyperic is graphing. The amount of data available for graphing from the Hyperic agent can only be described as bountiful. The agent returns detailed information for the processes and their attributes. For example, when monitoring a PostGres database, metrics are collected and graphed down to the availability, disk space, index space, rows per minute, scans per minute for each individual table! Hyperic does an excellent job of automatically collecting a great deal of information with little to no configuration required by the user. Ease of use is Hyperic’s strong point.

Many application servers, databases, and other system services are natively supported in the Hyperic HQ distribution, but if that doesn’t cut it, Hyperic offers a full-featured plug-in framework that extends its functionality.

Conclusion:
Hyperic is simple to configure. It has a large amount of information available for graphing, alerting, and a very well designed user interface that allows for easy navigation. One of the downsides to Hyperic is that there is a wealth of information, but graphing and saving this information isn’t as easily locatable to the user as it could be.

Although discovery is limited to servers containing agents, it provides a level of detailed information that is not available through other monitoring tools that use SNMP or SSH to collect information from hosts and devices. The graph displays in Hyperic are easily readable. Hyperic is full-featured, well-designed, and focused on performance and availability.

Hyperic HQ Community Edition is available here.

About the author:
Matthew Sacks works as a systems administrator. His focus is network, systems, and application performance tuning.