HQ 3.2 uses MySQL as a backend database, which makes HQ capable of monitoring and capturing up 1.5 million Web site metrics a minute. The metrics may be measures of response times by applications to Web site browser calls, length of response by Web site databases to queries, amount of memory used by applications, or speed of connection to other remote services.
Hyperic HQ previously supported its operations with the Oracle commercial database and open source PostgreSQL. Many of its customers already used MySQL, however, prompting the addition of MySQL support in the latest release, said Javier Soltero, CEO, in an interview.
Soltero said Hyperic seeks to make it possible to monitor a Web site, powered by thousands of servers, in real time and spot potential problems as they develop. Enabling Web sites to scale up their activity as traffic hits poses special problems for systems management. Hyperic is focused on "helping today's new generation of online companies maximize the availability of their large-scale Web infrastructures," said Soltero. In addition to MySQL, HQ 3.2 is also integrated with another piece of open source code, Nagios system monitoring. Nagios was an open source pioneer in system monitoring; other Web site management software, such as GroundWork, is based on Nagios. Hyperic now has tools to let users run Nagios functions in Hyperic and view Nagios Hosts, Statuses, and Services in the Hyperic user interface. Hyperic can chart and display results from many Nagios monitors and perform diagnostics on them, Soltero said. With all the information flowing into Hyperic, it's important to be able to do "management by exception," where the system recognizes departures from norms, instead of being swamped by a flood of data, he added.
For example, he said, if an application is running too slowly, the Hyperic administrator may view the Web servers that are hosting it. If they appear to be running OK, he can view the databases that are serving it, running a probe that can tell whether particular processes running against certain tables lead to a slowdown, and which ones are involved.
Release 3.2 "gives you more information and a greater ability to act on it," Soltero said.
Hyperic is capable of issuing about 15 reports, using open source JasperReports. In a Sept. 27 Information Week story, this figure was overstated as 35 reports, which can't be achieved without customer intervention and customization of some of the standard reports. Soltero said the number of reports will gradually increase with each successive release of the product.
Hyperic charges $500 a year per production server, regardless the number of processors, for an HQ 3.2 support subsciription. The price is cut in half for pre-production or test stage servers.