Today Pixie Labs emerged from stealth with its first product, Pixie. Pixie dramatically reduces the complexity and cost for developers to observe and troubleshoot their application’s performance. With Pixie, developers don’t need to change code, manually set up ad hoc dashboards or compromise on how much performance data they can observe. Early access customers range from early-stage startups to internet-scale streaming content teams who are all already running Pixie in their production Kubernetes environments. The product’s public beta starts today and will be showcased at Pixie Demo Hour on October 8: https://pixielabs.ai/events/pixieday.
Kelsey Hightower, Advisor to Pixie Labs, said: “All layers of software—application code, deployment infrastructure, network, OS, etc.—now expose APIs that allow us to treat infrastructure as data. Pixie helps us move to a world where we can freely access this data and focus on building data-driven workflows to observe, troubleshoot, secure and manage our applications.”
Having spent a decade building thinking machines, Pixie Labs’ founders Zain Asgar (CEO), an Adjunct Professor of Computer Science at Stanford University who was an engineering lead for Google AI, and Ishan Mukherjee (CPO), who led Apple’s Siri Knowledge Graph product team and was an early Amazon Robotics engineer, started Pixie Labs to empower developers with time-saving tools by intelligently augmenting their workflows.
Both had grown frustrated by how manual and inefficient it was to set up, manage and use existing systems to monitor distributed software applications. The decade-old monitoring and observability platforms require months of painstaking set-up and hours of manual data wrangling to troubleshoot when core business driving applications are on fire.
“We thought a better data system can be built, and we envisioned a magical developer experience where they can get the data they need without having to change code or move data outside their environments,” said Zain Asgar, co-founder and CEO of Pixie Labs.
As Pixie runs entirely inside a developer’s own Kubernetes platform, development teams are extending their established in-house or managed SaaS monitoring footprints to get instant and automatic visibility inside their very own production environments.
“Developers are superheroes without the data superpowers they need. We’re building Pixie to fill that gap. Pixie’s adoption in real production environments from fast-growing startups to internet-scale companies is starting to validate our focus on delivering a consumer grade experience to developers which respects and optimizes for their time,” said Ishan Mukherjee, co-founder and CPO of Pixie Labs.
Early access customers on Pixie:
- “I started seeing production data in Pixie after pasting one line to install. Amazingly, a few minutes into looking at the data, I was able to find and fix a faulty config that sped up our API requests by 14%. This is basically automatic Datadog for Kubernetes,” said Darshan Desai, founder and CEO of Plan, a productivity software company.
- “Real-time instant data access is a game changer. Insights that Pixie is able to deliver on what is happening within Kubernetes clusters is astonishing. Its use cases range from development to operational,” said Paulo Cabido, sr. director of engineering operations at ThousandEyes, a networking monitoring company now part of Cisco.
- “The Pixie engineers are giving instrumentation superpowers to the masses. As the complexity of hyperscalar style distributed services grows and engineers are fighting ever-more difficult bugs, the ability to add logging points to a running service brings debuggers to the 21st century,” said Andrei Matei, member of technical staff at Cockroach Labs, a database company.
- “Debugging and observability are two very important pieces in the software development lifecycle. Usually setting these up is painful and requires a lot of time. This is where Pixie makes life insanely simple. Right from a single line install to no instrumentation visibility, Pixie makes working with your applications and Kubernetes fun again!” said Roopak Venkatakrishnan, staff software engineer at Bolt, a fintech company.
- “Pixie is amazing. It seamlessly extracts monitoring data and enables users to perform live data dumps from applications running in Kubernetes clusters. All Kubernetes SREs should use it as part of their troubleshooting toolset,” said Mihai Todor, Lead Kubernetes SRE at Optum, a healthcare technology company part of UnitedHealth Group.
Pixie’s developer experience and customer adoption is driven by three fundamental technical breakthroughs:
- No-Instrumentation Data Collection: Pixie leverages novel technologies like eBPF (https://ebpf.io) to automatically collect baseline data (metrics, traces, logs and events) for the application, Kubernetes, OS and network layers. For last-mile custom data, developers can dynamically collect logs using eBPF or ingest existing telemetry.
- Script-based Analysis: Developers and operators use Pixie by running community contributed, team-specific or custom scripts from Pixie’s native debugging interfaces (web, mobile, terminal) or from integrations with established monitoring platforms. This code-based approach enables efficient analysis, collaboration and automation.
- Kubernetes Native Edge Compute: Pixie runs entirely inside Kubernetes as a distributed machine data system without customer data being transferred outside. This novel architecture provides customers a secure, cost-effective and scalable way to access unlimited data, deploy AI/ML models at source and set up streaming telemetry pipelines.
Today Pixie Labs also announced $9.15 million in Series A funding led by Benchmark with participation from GV.
“Since I was a monitoring product manager 20 years ago, the challenge with monitoring has been the same. First is how to get people to do the tedious task of instrumenting ahead of time. Second is how do people find signals in the overwhelming majority of monitoring data which is worthless,” said Eric Vishria, General Partner at Benchmark. “By using huge innovations in Kubernetes and eBPF and a clever new architecture which turns conventional monitoring approaches on their head, Pixie Labs has solved the two core challenges, giving people in-depth visibility with one line of code and only keeping the interesting data. I’m really excited about what becomes possible with Pixie.”
Additional Resources
- Read Zain Asgar and Ishan Mukherjee’s “Announcing Pixie’s Public Beta Launch and Series A” blog: https://blog.pixielabs.ai/blog/public-beta-launch/beta-launch
- Read Zain Asgar’s “Live Debug GoLang Applications in Production” blog: https://blog.pixielabs.ai/blog/ebpf-function-tracing/post
- View Pixie Labs’ “Observability on the Edge” deck: https://docsend.com/view/kj38d76
