MercadoLibre currently operates over 30,000 microservices running on 100,000 instances in Fury, its internal development platform. This scale, volume, and diversity of components creates a need for robust observability. But the massive amount of telemetry data that a production environment of this size generates creates new problems, such as noise and loss of cost control. For this reason, developing an observability ecosystem requires a strategy that resolves the needs of each of the different stakeholders within the organization and offers the necessary governance guarantees.In this session, the Observability team at MercadoLibre will share how they developed an enterprise-scale observability strategy that fits their dynamic hybrid cloud environment, with special focus on best practices on how you can use custom metrics to enhance visibility and streamline incident management—all while keeping costs in check.