K8S Monitoring, Application Debug with Pixie

Introduction

3 min readApr 23, 2023

I have been working professionally in the software industry for about 15 years. I have had the opportunity to work with many different tools in different companies, such as Appdynamics, Dynatrace, New Relic, Datadog etc. Thanks to these applications, we can observe our application performance metrics, traces and bottlenecks. Today I will explain to you how to set up and manage another observability tool (plus it’s open-source), Pixie!

>> This article was originally published on my blog:
https://habil.dev/kubernetes-monitoring-application-debug-with-pixie/

What is Pixie?

Pixie is an open-source observability tool for Kubernetes applications. Pixie uses eBPF to automatically capture telemetry data without the need for manual instrumentation.

Thanks to Pixie we can view the high-level state of our clusters.

service maps
cluster resources
application traffic and more

Also drill down into more detailed views (pod state, flame graphs, individual full body application requests etc)

Pixie automatically collects the following data:

Protocol traces
Resource metrics
Network metrics
JVM metrics
Application CPU profiles

Install

To set up Pixie, you need to sign up first. After account creation, you need to log in with the px auth login command and decide on the installation method. You can install Pixie via CLI, Helm or Yaml. I'll go with CLI.

bash -c "$(curl -fsSL https://withpixie.ai/install.sh)"

brew install pixie

Deploy

Pixie is using Operator Framework, which is another important point for the deployment. If you already have it you need to add some flag to the install command.

No OLM present on the cluster:

px deploy

OLM already exists on the cluster:

px deploy --deploy_olm=false

After the deploy command is executed, Pixie will create 3 following namespace:

olm
pl
px-operator

💡Pixie stores all the data only in your cluster.

For a quick test after a successful deployment, you can run the command below

px live px/http_data

This pre-defined command will give you detail about HTTP data.

If you wonder about all script lists, you can execute the following script:

px scripts list

Web GUI

Thanks to the Web UI, you can get live information about your cluster and applications.

Before starting, you can access all the scripts that you can access with the CLI by selecting the namespace, also via the UI.

Service Map:

K8S Nodes:

Pods:

Also, you can find about JVM, the slowest SQL queries in your dashboard.

Conclusion

In this article, we learned how to install Pixie and the basic commands. Now we can access more information about our applications and take precautions by observing before problems arise.

See you in the next article. 👻