The ability to integrate SLA (service-level agreement) monitoring, analysis and reporting makes ProactiveNet 6.0 appealing to a new group of stakeholders: "senior-level IT management and business unit management," said David Langlais, vice president of marketing at ProactiveNet in Santa Clara, Calif.
"By looking at the business level, they can see, How much revenue or productivity am I getting from my applications?" he said.
Increasing pressure on IT to run its function as a service has fueled demand for SLA management, although only a handful of tools provide such functionality. ProactiveNets update gives customers the ability to create and manage SLAs on a more granular basis and then roll up specific SLAs to view overall performance.
"You can set an SLA value for any attribute in the system, and you can create hierarchical SLAs," Langlais said. Those can include SLAs for software, application servers, network devices and transactions, and all of them can be rolled "into a single performance SLA for online banking."
Once an SLA violation occurs, operators can drill down through the hierarchy to learn what the problem is and which sub-SLAs are causing it. "You can directly link SLA noncompliance into the problem in the IT infrastructure causing the violation," Langlais said.
ProactiveNet 6.0 also adds the ability to monitor several servers that are set up in a cluster and linked to a load balancer as a single logical entity. This new resource-pools monitor can reduce the number of events generated because it looks only at a single threshold, but it still provides visibility into individual components in the pool.
"You can click down through it and find out which subcomponent is causing a problem," Langlais said.
ProactiveNet is a self-learning tool that baselines the performance of all of the attributes it learns about to determine each attributes normal operating environment. Intelligent thresholds follow the performance of each attribute, whether its CPU utilization, the number of SQL write requests waiting or the size of the Java bean pool on an application server. The software then looks for abnormalities rather than sifting through thousands of events.
The new release also improves the precision of ProactiveNets root-cause analysis by better understanding the relationship between infrastructure components and applications.
"We work at the basis of groups and relationships with our correlation," Langlais said. "Since we ignore everything thats normal and focus on performance abnormalities related to the infrastructure associated with the SLA, we understand vertically how different pieces of a Web server are related.
"We can ignore everybody thats not in [a specified] group. Were filtering down and excluding less-likely suspects."
The release is due next week. The new SLA module will be priced separately. A starter pack that includes the module and 40 or 50 predefined SLAs—enough for two applications—will cost about $30,000.