Pixie Labs Inc. exited stealth mode today with $9.15 million in funding from Benchmark and GV, Alphabet Inc.’s venture capital arm, to change how developers monitor their applications.

Application monitoring is an important part of enterprise software projects. Development teams need visibility into the inner workings of their software to ensure that it’s functioning as expected and to perform troubleshooting when users report an issue. Usually, monitoring is performed by integrating special data collection code into an application that gathers diagnostics information.

Pixie says its namesake cloud-native monitoring tool collects diagnostics information without requiring extra code to be added to applications. Eliminating this step, according to the startup, has the potential to save up to months of work. Reducing the number of moving parts in a software project has the added benefit of reducing code complexity, which can translate to fewer bugs and security issues.

Pixie’s monitoring tool is delivered as a collection of integrated services that developers can install on Kubernetes. The software collects diagnostics information without embedded data collection code by using a feature called Extended BPF in Linux. Linux is the operating system with which Kubernetes is most often used in the enterprise. 

The Extended BPF feature opens up access to the so-called kernel memory, a normally off-limits section of a server’s RAM that is reserved for internal operating system components. Inside the kernel memory, Pixie intercepts the machine data that is generated from applications running atop Linux and then exports it. Developers can then integrate this data into their monitoring workflows. 

“Pixie leverages novel technologies like eBPF to automatically collect baseline data (metrics, traces, logs and events) for the application, Kubernetes, OS and network layers,” Pixie co-founders Zain Asgar (pictured, right) and Ishan Mukherjee (left) wrote in a blog post today announcing the launch of the startup’s tool into public beta. Asgar, the startup’s chief executive, is an adjunct computer science professor at Stanford University, and Mukherjee previously led Apple Inc.’s Siri Knowledge Graph product team.

On top of its core data collection capability, Pixie provides a programming language for processing the incoming information. Developers can write scripts to perform tasks such as isolating events of interest or identifying which services are exchanging data with one another inside a Kubernetes cluster. 

Pixie’s tool is freely available. The startup plans to make money by launching a paid version for enterprises further down the road. The tool has already started gaining some market traction, winning over early adopters at companies such as heavily funded database startup Cockroach Labs Inc. and Cisco Systems Inc.’s ThousandEyes subsidiary, among others. 

Photo: Pixie Labs