Date
Time
-
Location
TBD
What happens when many teams instrument their systems independently? At leboncoin —part of the Adevinta group— we faced a shattered monitoring landscape. With multiple marketplaces across Europe, various technology stacks and a bunch of different naming conventions, answering simple questions like "What's our service error rate per marketplace?" turned into a wild hunt through custom metrics, custom dashboards, redundant implementations, and siloed knowledge.
In this talk, we’ll walk through how we brought order to the chaos by adopting OpenTelemetry's semantic conventions. We'll detail how we aligned engineering teams around common metric nomenclature, explain why OpenTelemetry emerged as our go-to solution, and how we plugged everything with Datadog, as a unified platform, to build dashboards and alerts that work for everyone, to observe the services we provide to our dozen of millions of unique users each month.
We'll also share the challenges we faced: securing stakeholder buy-in for migration, maintaining operational visibility during the transition, and the incidents that became valuable learning opportunities.
Finally, we'll give you a glimpse of what's next — our journey towards observability-as-code and how it's reshaping the way we monitor at scale.