There are probably not too many of you with your hands still in the air--and with good reason. Big Brother was innovative and, for its day, a godsend. But today's complex Web infrastructures and intense customer requirements for speed and quality require a much more comprehensive solution--one that comprises more than just a ping check to determine availability and performance.
Today's Web monitoring solutions include a wide variety of checks and balances that provide not only performance and availability statistics, but also comprehensive details about the disparate pieces of hardware and software that make up your Web site and keep it available and running in top form.
The biggest problem with these services is the almost overwhelming number of them available. How do you choose one over the other? That's where we come in. We'll discuss how to choose the Web performance-monitoring service that fits your needs.
Protocols
First, ascertain if the service supports the protocols necessary to report on the performance and availability of the distinct pieces of your Web site accurately. DNS and HTTP should be a given, with HTTPS (HTTP Secure) following as a close third. If you're providing e-mail, you'll want the ability to monitor SMTP and POP3/IMAP, and, of course, you'll want to ensure that your FTP site is available too. There aren't many services out there that don't offer support for the necessary protocols, so simply make this a checklist item and mark it off.
FreshWater Software SiteSeer offers e-mail protocol support in addition to the minimum required support for HTTP, HTTPS and DNS, but Keynote Systems' WebSite Perspective, for example, supports only HTTP, HTTPS and Flash. Mercury Interactive Corp.'s ActiveWatch not only offers DNS, HTTP and HTTPS, but provides support for streaming media, WAP and e-mail protocols.
Content Verification
Knowing that your site is available with just a ping or even a TCP open-and-close simply isn't enough anymore. Customers are picky: They want their pages to appear quickly and accurately (go figure!). Content verification provides assurances that the site is correct from a functional and content perspective--both of which are just as important as the speed with which the pages load.
The ability to verify content goes deeper than a simple check for certain words in an HTML page. Make certain the service you're choosing can provide verification of all aspects of your Web site, including Java applets, Flash content and streaming media. If it's part of your Web site, you'll need to make sure it's available and correct. Mercury Interactive's ActiveWatch offers, by far, the most comprehensive content-verification support.
Although all the products listed in our Interactive Buyer's Guide charts can provide verification of HTML, only Mercury offers verification for Flash, Java and streaming media. If your site contains these elements, consider your choice of services very carefully. If you have faith in your quality-assurance process or monitor content internally for accuracy, a service without a comprehensive content-verification system may suffice. But make sure your internal processes are accurate and acceptable before you sign up for a service that doesn't support these components.
The Data
When problems arise, you want to know exactly what went wrong. The service you choose should provide detailed information about the performance of each component on your Web site. If DNS is the problem, the data reported by the service should indicate as such. If the problem is a sudden increase in network latency, you'll want to be able to see this at a glance through a summary graph provided by the service. But notification is only half the game; enabling root-cause analysis is a must for a successful Web monitoring strategy. Any service should assist you in this process by accurately providing a granular view of performance statistics.
ActiveWatch offers im-pressive integration between your Web site performance statistics and data collected via the product's Topaz WeatherMap, which gathers performance stats on the Internet backbone itself. Such an integrated system can enable root-cause analysis down to a single peering point or NAP (network access point) on the Internet backbone, making your job that much easier.
Other providers, including Atesto Technologies with its Performance Suite 2.0, BMC Software (SiteAngel) and Empirix (FarSight), offer inside-the-firewall companion products that provide correlated views combining infrastructure performance with end-user experience. This gives you a deeper look into root-cause analysis of performance issues. Another provider to watch is Segue Software, which sells inside-the-firewall monitoring solutions and will soon have a hosted version of its solution. Segue says it will integrate this service with its current software for a hybrid monitoring solution.
In this same vein, consider how well services with behind-the-firewall solutions use testing scripts in their hosted monitoring products. If you've already built testing scripts in-house to validate the data, it's a boon if the service will use those same scripts to perform its monitoring tests. This ensures not only that you're gathering data on your system's performance but that the data gathered represents a scenario simulating an actual user's experience.
To analyze performance trends, also think about how long gathered data is stored. Examine both how long the data is available online, as well as how data is handled after certain periods of time. Some services, such as BMC's SiteAngel, provide data for more than one year. But after a few months, SiteAngel will compress the data into more generalized reports that leave much of the granularity of the original data out. Most services, including ActiveWatch, Performance Suite 2.0 and FarSight, maintain performance data for weeks or months.