Zalando’s application landscape has evolved over the past decade from being a PHP/Java based monolith into a highly distributed microservice architecture with thousands of moving parts. Operating this environment brings huge challenges in understanding system health, enabling operational visibility, routing alerts and facilitating effective incident response. Even conceptually simple questions like "Which applications are running in production?" can develop a surprising depth.

In the talk we are going to provide a high-level overview about the telemetry systems that Zalando is using in order to operate its application landscape and how they evolved over the past 5 years. In particular we will cover how we rolled out Distributed Tracing across the company and scaled our home-grown metrics system to handle 100M active metrics with a vendor provided TSDB backend.