What is Pixie used for?

Pixie is an open-source Kubernetes observability tool that uses eBPF to gather telemetry automatically and stores the data inside your cluster. It is a good fit for debugging microservices on Kubernetes, profiling application performance in a cluster and observability with no code instrumentation.

Is Pixie open source?

Yes. Pixie is open source under the Apache-2.0 licence. Its source code is on GitHub at pixie-io/pixie and is written mainly in C++.

Yes. Pixie is open source, so the software itself is free to use.

Can I self-host Pixie?

Yes. Pixie can be self-hosted on your own server or infrastructure.

What is Pixie an alternative to?

Pixie is an open-source alternative to Datadog, IBM Instana and Groundcover. Other open-source alternatives to Datadog include Coroot, Odigos and SigNoz.

Is Pixie actively maintained?

Yes. The most recent commit to Pixie was on 24 September 2026. The project has 6.5k stars on GitHub.

Pixie: open-source alternative to Datadog, IBM Instana

About Pixie

Pixie is an open-source observability platform made for applications running on Kubernetes. It gives a high-level view of a cluster, including service maps, resource usage and application traffic, and lets you drill into individual pods, flame graphs and full request bodies without adding instrumentation code to your services.

It relies on eBPF to gather telemetry automatically, covering requests, resource and network metrics and application profiles. Data is collected, stored and queried on the cluster itself, an approach the project calls in-cluster edge compute, which keeps raw telemetry local and limits overhead. Scripts written in a Python-based, dataframe-style language let you build custom views, and the project is hosted by the Cloud Native Computing Foundation.

Pixie is licensed under Apache-2.0 and is installed into your own Kubernetes cluster, whether a local one such as minikube or a managed cluster on GKE, EKS or AKS. It suits platform and application teams that want fast insight into service behavior without a separate agent-heavy monitoring stack.

Key features

Automatic eBPF-based telemetry collection
Service maps and cluster resource views
Flame graphs and application profiling
Full-body request inspection per pod
In-cluster data storage and querying
Scriptable custom views with Pixie scripts

Good fit for

→Debugging microservices on Kubernetes
→Profiling application performance in a cluster
→Observability with no code instrumentation

Built with: C++; Go; Kubernetes
Tags: kubernetes; observability; ebpf; monitoring; cncf; profiling; self-hosted

Pixie: questions and answers

What is Pixie used for?: Pixie is an open-source Kubernetes observability tool that uses eBPF to gather telemetry automatically and stores the data inside your cluster. It is a good fit for debugging microservices on Kubernetes, profiling application performance in a cluster and observability with no code instrumentation.
Is Pixie open source?: Yes. Pixie is open source under the Apache-2.0 licence. Its source code is on GitHub at pixie-io/pixie and is written mainly in C++.
Is Pixie free?: Yes. Pixie is open source, so the software itself is free to use.
Can I self-host Pixie?: Yes. Pixie can be self-hosted on your own server or infrastructure.
What is Pixie an alternative to?: Pixie is an open-source alternative to Datadog, IBM Instana and Groundcover. Other open-source alternatives to Datadog include Coroot, Odigos and SigNoz.
Is Pixie actively maintained?: Yes. The most recent commit to Pixie was on 24 September 2026. The project has 6.5k stars on GitHub.