Sampling is a critical tool in observability systems since it allows us to lower the cost of generating, collecting, and analyzing telemetry data in our systems. It is a form of finding the balance between costs and performance, but the road to success is paved with obstacles.

In this session, you will learn why and how to collect the most relevant and insightful data rather than the most common. We will look into how Head and Tail sampling work and explore how to configure them to collect the most crucial data to handle scale in production. We will dive into sampling overuse use cases, learn how to avoid them, and manage cost.

You will leave this session with a complete playbook on sampling, so you're ready to implement everything you've learned immediately.