Product Reviews

An Event Monitoring Solution for your Network

Argent Guardian offers reliable, comprehensive and flexible event monitoring.

• By Jim Richards
• 09/01/2001

Ask a dozen people, "What makes an event monitoring package great," and you may get a dozen answers; but everyone would probably agree that reliable, comprehensive and flexible event monitoring and notification is a must. The product must be easy to use, have great built-in reporting tools, should be designed on a stable platform, and shouldn't cost you your first born to own.

Welcome to the world of The Argent Guardian. In a world of half-million-dollar event monitoring solutions like Tivoli and NetIQ, Guardian steps in at a far lower cost.

Guardian is built on top of the job engine used in the company's Job Scheduler software. This is what makes it so great. Almost every object within the program has some sort of calendaring feature built into it. Guardian has the flexibility to manage exactly how and when your jobs will run. You can configure when to exclude jobs for maintenance, how often jobs will send alerts, how quickly and how often the escalation process happens, and even when to use a specific alert transport like dial paging vs. e-mail. The core of this flexibility lies in highly extensible calendar definitions called "every" definitions. These allow you to define a customized variable of dates yearly, monthly, quarterly and so on. These "every" definitions then get combined into "base" definitions for use in jobs, alerts and other runtime features of the main application. Better yet, everything is manageable either at the server, on a remote management station or over the Web.

If you can use Windows Explorer, you can use The Argent Guardian. (Click image to view larger version.)

The types of alerts the system can fire are almost limitless. Guardian can alert via a simple system beep, send a page, or run a pre-defined script then send a page and beep the system. For example, suppose you wanted to check your SQL logs for a specific string that appears within a period of time. If the string is found, Guardian can be told to execute a SQL-stored procedure and then send an e-mail notifying the on-call SQL administrator that the issue has been resolved. The jobs can even be set up so that other criteria have to be met before a job is allowed to execute an alert.

The main type of data the program collects is native Windows information, which doesn't have a large impact on system resources. This is nice because it means that the program isn't in competition for server resources with your user applications. The program gives you the option of using client agents, which is a blessing in the world of Microsoft. I've never met an administrator (NT, SQL or otherwise) who said they liked agents. As far as I'm concerned, agents are just more trash on a server that can interfere with critical network services. If you need to connect through a firewall, however, you can use the product's firewall agent (basically a proxy-based service) to forward event information on behalf of the monitored client.

Guardian was fairly easy to implement, too. I put the package into full-production monitoring on 80 of our servers in less than two weeks. This included event monitoring on all critical network services and collecting uptime metrics that are published on our intranet via the product's built-in reporting and graph capabilities. I found the product to be highly responsive, allowing me to change and modify what and how I was monitoring on the fly. The GUI is clean and crisp, making it easy to use. There isn't a lot of clutter to sift through.

Guardian's features are intuitive, and the product includes many demos to help you figure out what they're intended to do. There are great testing and tracing capabilities built in so you can quickly check portions of the system to diagnose anything that isn't functioning. The package currently ships with 23 monitoring plug-in rule components included. The plug-ins include rules for Exchange, SQL, Oracle, APC, Cisco, Nortel and many others. The system can also capture SNMP traps and will integrate with Compaq Insight Manager and Tivoli.

The one thing I didn't like about Guardian is the frequency of new releases (almost one per month). This caused a lot of problems with the QC process and throughout my testing. I worked with the Guardian development team to identify and fix bugs that could've been discovered with proper regression testing. The good news is that the whole Guardian team is blessed with an above-and-beyond mentality. The tech support is outstanding (they actually call you back); if you don't get what you're after they'll send you right to the main developer.

The more I dove into this system, the more I liked it. If you're looking for a system that will notify you when you want, about exactly what you want, and take only the actions you want, you must check out Guardian. You can download a 30-day evaluation version from its Web site at www.argent.com/products/download.cgi?product=monitor. This is a full-release version that will monitor up to 100 servers.