Year-old start-up company AppFirst is trying to shake up the way IT administrators think not only about application performance management but also server monitoring in general.
The company offers a proactive monitoring service that uses a patent-pending technique for collecting performance data off of servers in real time, set at 20-second intervals by default but customizable, says Pamela Roussos, chief marketing officer at AppFirst. This compares to traditional poll-based server monitoring products that collect data at intervals of five minutes or longer.
"We won't miss any incident," she says.
With the AppFirst service, users download collectors for use on servers no matter whether they are physical, virtual or in the cloud. Sitting at the system call level, they capture and record all activity transpiring between the application and the operating system in real time. The overhead is almost negligible, Roussos says -- 0.6% on Linux systems and 1% on Windows servers.
"We watch running applications as they shoot across the entire application stack, and do that in a way that won't affect performance and agnostic to the language, components and databases in use," she explains.
What’s more, AppFirst inherits the output of data gathered via Nagios, the open source IT monitoring software, as well as the thousands of Microsoft-created Windows performance counters, Roussos says. It installs 50 Nagios plug-ins, and automatically will pick up output from any other of the 1,800 or so such plug-ins, she adds.
"We correlate this data with our information to provide a complete picture of what's going on in the entire application stack," she says. "Because we have a complete data set, we can see change happening easily and quickly and can alert to those changes immediately."
All of this, AppFirst believes, should be free to IT administrators. "What we figured is why should you be paying for things that aren’t really even serving your needs?" Roussos says.
Yes, IT administrators need to know when things are going amiss -- and that they can now get from the company's fre AppFirst Basic SaaS offering, which monitors servers and alerts upon problem discovery. But what they really need is a way to get to the root cause -- quickly and easily, she says.
And that AppFirst does via its fee-based Professional version, which provides deterministic root-cause analysis. "We can tell users what's going on in the server, with what process and what that process is doing to affect the performance of an application," Roussos says.
If IT administrators only need server monitoring, they’re free to use the Basic version for as long as they want. But the minute they need help with root-cause analysis, they can switch to the Professional version via an easy account status change in their management consoles and credit-card entry, she says.
Priced on a per-usage basis, AppFirst Professional costs 2.5 cents per hour, which translates to $18 per server monthly for constant 30-day use. With a 75-server bundle, that monthly price comes down to roughly $10 per server, adds Roussos, nothing that custom plans also are available for larger enterprises.
"AppFirst is filling two gaps in the market -- one is in the underserved SMB area, where there are a lot of point tools that mostly do network monitoring or higher-end solutions [that are too costly,] and the other is for companies trying to control the overhead of managing application performance,” says Bojan Simic, president of Trac Research.
"Between the combination of its capabilities, the price point and ease of use," he says, "AppFirst makes for an attractive application performance management offering."
Read more about infrastructure management in Network World's Infrastructure Management section.