Master Continuous Profiling on Kubernetes with Pyroscope: A Step‑by‑Step Guide

Developers often need to locate performance bottlenecks in production applications. Traditional log‑based debugging is time‑consuming and lacks detail. A modern approach is to use continuous profiling tools that highlight the slowest code paths consuming the most resources.

What is Pyroscope?

Pyroscope is an open‑source continuous profiling platform consisting of a server and agents. It efficiently collects, stores, and queries profiling data using a purpose‑built storage engine (Badger) and supports language‑specific agents for Go, Python, Ruby, eBPF, Java, .NET, PHP, and Rust.

Continuous Profiling

Continuous profilers run in production, gathering CPU and memory usage over time. They quickly reveal resource‑intensive functions after new code is deployed, helping reduce cloud costs and improve user experience.

Available Continuous Profilers

Pyroscope

Parca – eBPF‑based sampler for C, C++, Go, etc.

Datadog Continuous Profiler

Google Cloud Profiler

Why Use Pyroscope?

Compared with Datadog and Google Cloud Profiler, Pyroscope offers a low‑overhead storage engine, language‑specific agents, and eBPF support, providing more readable profiles with minimal CPU impact.

Installation on Kubernetes

Start a local Kubernetes cluster with Minikube: minikube start Add the Pyroscope Helm chart repository and install the chart:

helm repo add pyroscope-io https://pyroscope-io.github.io/helm-chart

helm install pyroscope pyroscope-io/pyroscope --set service.type=NodePort

Verify the installation:

helm list

kubectl get all

Integrating Pyroscope with Sample Services

Modify Dockerfiles to include the Pyroscope binary and launch the application through the profiler.

Python Email Service

COPY --from=pyroscope/pyroscope:latest /usr/bin/pyroscope /usr/bin/pyroscope
CMD [ "pyroscope", "exec", "python", "email_server.py" ]

Build and push the image:

docker build . -t beellzrocks/emailservice:latest
docker push beellzrocks/emailservice:latest

.NET Cart Service

COPY --from=pyroscope/pyroscope:latest /usr/bin/pyroscope /usr/bin/pyroscope
ENTRYPOINT ["pyroscope", "exec", "-spy-name", "dotnetspy", "/app/cartservice"]

Build and push the image similarly.

Go Product Catalog Service

import (
    pyroscope "github.com/pyroscope-io/pyroscope/pkg/agent/profiler"
)

func main() {
    pyroscope.Start(pyroscope.Config{
        ApplicationName: os.Getenv("APPLICATION_NAME"),
        ServerAddress:   os.Getenv("SERVER_ADDRESS"),
    })
    // code here
}

Build and push the Go image after adding the profiler start call.

Updating Kubernetes Manifests

Replace container images with the newly built ones and add the required environment variables and security context.

containers:
  - name: server
    image: beellzrocks/emailservice
    env:
      - name: PYROSCOPE_SERVER_ADDRESS
        value: "http://pyroscope:4040"
      - name: PYROSCOPE_APPLICATION_NAME
        value: "email.service"
    securityContext:
      capabilities:
        add:
        - SYS_PTRACE

Apply the manifests:

kubectl apply -f https://raw.githubusercontent.com/infracloudio/microservices-demo-dev/master/release/kubernetes-manifests.yaml

Expose the Pyroscope UI: minikube service pyroscope Open the displayed URL (e.g., http://192.168.49.2:30639) to view profiling data.

Resource Utilization

Pyroscope itself consumes very little CPU and memory while storing data locally in Badger. Screenshots show low CPU usage and efficient storage.

Flamegraph Examples

Flamegraphs for the Go product catalog, .NET cart, and Python email services illustrate how Pyroscope visualizes hot code paths per language.

Conclusion

Continuous profiling is essential for meeting end‑user performance expectations. Pyroscope provides low‑overhead, language‑aware profiling that helps developers diagnose issues early, optimize code, and reduce cloud infrastructure costs.