As Zendesk migrated from a monolithic application to an ecosystem of services, being able to reliably monitor the interactions between those services became crucial. From the early days, teams have been innovating on observability and sharing their best practices with other engineering teams.

As the company grew from a handful to now hundreds of production services, and to dozens of different engineering teams, the need for a more fully unified and standardized observability became a chief concern.

In this talk, Daniel Schierbeck will share how adopting a service mesh has helped Zendesk manage their growing number of services, while allowing them to standardize on their observability. He will also explain the evolution of how they have approached monitoring service interactions at Zendesk, from the early days of primitive logging, to their adoption of Datadog metrics and then Datadog APM—and where the adoption of a service mesh fit into the whole picture.