Monitoring
The
proliferation of computers and services in horizontally scalable systems
or grid computing complexes poses a problem for the enterprise CIO: although
deployment to racks of commodity computers decreases procurement costs,
the costs of management and monitoring those computers can skyrocket.
And maintaining adequate service levels can be difficult when hundreds
or thousands of servers are used. There are many choices available to
the CIO for management and monitoring tools, both commercial software
and Open Source.
The CollabWorks Monitoring Project recognizes that the many monitoring tools available to the enterprise are difficult to select, customize, and manage. None of the tools work optimally in all possible environments. Selecting multiple tools poses interoperability and integration problems, for the systems and applications being managed, as well as for each other.
The Project Canary monitoring tool is a flexible framework originally developed by Sun Microsystems. Canary is script-based, lightweight, portable across many operating systems, easy to install and manage, and provides rich historical usage information in an easy to use graphical format.
The Canary framework can track network traffic, CPU utilization, memory and disk utilization, and application profiles. Its many graphical output formats can show runaway processes and unusual data patterns. By examining historical trends, administrators can identify anomalies, determine root causes of problems, and correct them.
As the Project started to analyze member requirements, however, the scope of work was expanded to include other monitoring frameworks, such as NagiosŪ and Ganglia. Each of the frameworks has its own strengths and weaknesses, and understanding how they relate to different system, network, and application configurations is an essential part of creating a services monitoring strategy.
The Project is beginning its work by developing a series of scripts that enterprises can run to find what monitoring frameworks are already in place. Anecdotal evidence indicates that there are many monitoring tools running unsupervised and unknown to the CIO, a consequence of employees who experiment with tools, or leave the enterprise without informing anyone of what they have installed. This can lead to wasteful system and network utilization and adds no value to the enterprise CIO.
Activities of the Open IT Monitoring Project include:
- A collaborative environment in which enterprise IT systems monitoring experts can interact with their peers and can invite external consultants and experts to participate
- Discovery of enterprise monitoring usage.
- Shared assessment of tools and scripts to enhance monitoring frameworks.
- A database of monitoring frameworks, candid member reviews of their experiences, and evaluations of their interactions with popular software products.
- Best practices and policies for interpreting and reacting to monitoring reports.
- Enhancing popular open source monitoring framework to meet common enterprise requirements, such as measuring unique characteristics of operating environments, or integrating with applications, services, and other popular monitoring frameworks.
- Confidential sharing of monitoring reports, allowing enterprises that have common application characteristics to establish baseline expectations of operations and performance.
The collaborative approach to enhancing monitoring frameworks and sharing best practices service analyses can reduce enterprise IT expenses, allowing funds to be spent on more productive, innovative, and differentiated activities.
*Nagios and the Nagios logo are servicemarks, trademarks, registered servicemarks, or registered trademarks owned by or licensed to Ethan Galstad.
