Pixie-demos is a collection of demonstration projects that showcase observability and monitoring capabilities for Kubernetes environments using eBPF-based instrumentation.
The repository contains multiple self-contained demo projects that illustrate how to solve specific operational challenges in Kubernetes clusters. These demos cover use cases including canary analysis with Argo Rollouts, horizontal pod autoscaling based on custom metrics, detecting unauthorized processes like cryptocurrency miners, continuous CPU profiling, API endpoint deprecation analysis, garbage collector instrumentation, HTTP/2 traffic tracing, cluster cost estimation, SSL/TLS connection tracing, and OpenTelemetry data export. Each demo pairs working code with accompanying documentation to show both the problem being solved and the implementation approach.
The demos are designed for teams operating Kubernetes clusters who want to understand how to apply eBPF-based observability to real operational problems. They work best for organizations evaluating whether Pixie fits their monitoring needs or looking for concrete examples of how to implement specific observability patterns. The projects span infrastructure concerns like cost tracking and security monitoring, as well as application-level concerns like performance profiling and API lifecycle management. These are not tutorials for Pixie itself but rather demonstrations of what becomes possible when combining Pixie's observability capabilities with Kubernetes-native tools like Argo Rollouts and the Kubernetes metrics API.
The repository maintains a diverse set of working examples across different technical domains, from low-level kernel instrumentation with eBPF uprobes to higher-level Kubernetes integration patterns. Each demo is independently deployable and includes the necessary configuration and setup instructions. The project serves as a living reference for practical applications of observability in containerized environments.
