A monitoring tool that scales from minuscule to mammoth networks

Despite Windows 2000's much-touted reliability, network administrators can still run into problems. Although blaming the OS is easy, external variables such as device failures and overloaded systems cause many network problems. If you've ever had trouble tracking down a network problem, Argent Software's The Argent Guardian is the tool for you. The product is a scalable monitoring solution for Windows NT and Linux networks. For once, the marketing rhetoric is true—The Argent Guardian scales to handle both the smallest and largest of networks.

After I downloaded and registered the software, I quickly configured the program to monitor my network's crucial aspects. As Figure 1 shows, the main console uses a simple tree-based user interface (UI), so administrators can configure and use the program without learning a new UI paradigm. However, the console is a Win32 program, and the product lacks an HTML-based interface. This deficiency limits an administrator's ability to configure and maintain The Argent Guardian on non-Windows platforms.

One of The Argent Guardian's unique advantages is that the monitoring architecture doesn't require you to install agent services on monitored workstations. The primary advantage of this type of system is that you can monitor your secure client systems without introducing foreign software into the environment.

Configuring Rules
The monitoring engine uses a flexible rules-based design. The software's rule set lets you define how the system detects problems. The rule set consists of eight classes: event, performance, program, service, SNMP trap, command, system down, and printer. You use this class order for grouping purposes. Instead of learning another scripting convention, you simply select the appropriate rules within each class to construct a working rule set.

After you establish the rules, you can select the notification methods that The Argent Guardian uses to contact you in the event of a problem. The product supports all the common notification methods, such as pager alerts, email messages, network messages, SNMP traps, and system reboots. You can also schedule batch jobs that run when something goes awry.

After you configure the rules and alerts, you need to define a node list. In Argent terminology, node lists are groups of servers that share a common characteristic. For example, you might group your application servers or servers that reside in a particular geographic location.

To simplify your use of rules, notifications, and node lists, the product includes support for relators, reusable templates that you can apply to multiple events. Relators let you bind all three data object types together. Finally, to set The Argent Guardian into motion, you use Argent's scheduling facility to coordinate recurring monitoring events.

The Test Drive
For my testing, I created a simple rule set to ensure that a remote server running Win2K Advanced Server was functioning. This server creates daily backups of crucial data to optical media. The backup job is a quick scripting hack that runs my backup utility, then notifies me when the process is complete. Monitoring the backup job was simple: The Argent Guardian includes a rule that scans the event log for an End Backup entry. However, to make the system act as an application server, I set it up as a Win2K Server Terminal Services and Citrix MetaFrame hybrid. Therefore, the machine was working harder than a dedicated backup server typically would. I needed The Argent Guardian to ensure not only that the server was operational but also that the system could sufficiently service my users, properly run the backup job, and keep out unauthorized users. In short, I needed the product to perform the job of a network administrator.

I used The Argent Guardian to define performance rules for CPU time usage, hard disk space consumption, and memory utilization by users running applications off the server. If the system encountered two of four criteria (i.e., if the CPU time exceeded 90 percent usage, if free disk space fell below 20 percent, if the system hit the pagefile too often, and if the network load became too heavy), the performance rules would trigger an alert message. Because Microsoft IIS was driving my MetaFrame system, I also configured the product to monitor my Web server's cache misses and security failures. As I expected, these problems occurred. The Argent Guardian sent an alphanumeric page coupled with an email message, which I picked up using a Wireless Application Protocol (WAP)-enabled cellular phone. The product directed my attention to the server's memory situation: Performance was slowing to a crawl because the server didn't have enough physical memory to service multiple users. The backup task was still running smoothly, but The Argent Guardian's immediate notification—regardless of my location—was comforting.

The Argent Guardian's design and scalability features are impressive whether you have a small three-system workgroup, a large multisystem domain, or a huge WAN spanning three sites and two coasts. The product uses one architecture, so it will grow with your network.

If your company handles any type of e-commerce, The Argent Sentinel program (included with The Argent Guardian) will justify the cost of the entire package. The Argent Sentinel is a Web-monitoring program that lets you keep an eye on your competitors. Argent divides the program according to the functionality that each applet provides. You use The Argent Sentinel's Web Site Change Monitor to watch your competitors' Web sites for price changes that might affect your e-commerce site. For example, you might configure the monitor so that price changes on specified items trigger alert messages to the sales department.

The Argent Sentinel's Web Site Response Monitor, which Figure 2 shows, tests Web site performance by downloading several pages from a designated site. Although the tool is great in concept, the nature of the Internet hinders its effectiveness. External variables such as latency and poor router configuration can grossly misrepresent a Web server's actual performance. However, this tool can accurately measure your internal Web server's performance.

In the same vein, The Argent Sentinel's Web Site Reliability Monitor lets you perform a thorough examination of your Web site. Missing links and images are common on complex Web sites; this monitor keeps your Web site looking and running as you designed it. Working with the Web Site Reliability Monitor is the E-Mail Performance Monitor, which triggers an alert when an email message takes too long to reach its recipient. Finally, the Check FTP Site Performance Monitor lets you test your corporation's FTP site performance.

Because Argent directly integrates The Argent Sentinel with The Argent Guardian, you can use The Argent Sentinel with your relators to monitor your entire network operation. You don't need to use a separate program to monitor your Internet servers, so you'll cut down on cost and administration time.

A Compelling Package
In The Argent Guardian, Argent combines a small footprint with extreme scalability that is limited only by your hardware. An intuitive UI lets you easily define and customize rule sets, and The Argent Sentinel component alone is compelling enough to justify the package's cost. Full phone support is available 24 * 7. You can download The Argent Guardian from the company's Web site.

The Argent Guardian 4.0A
Contact: Argent Software * 860-489-5553
Web: http://www.argent.com/
Price: $9000 for the first 10 servers, $8000 for each additional 10-server group, volume discounts available
Decision Summary:
Pros: Extremely scalable; doesn't require agent services on client systems; lets you monitor locked-down computers
Cons: Monitors only Windows-based systems