'' A resilient system relies heavily on observability, particularly at the record level, to identify failures and understand why they occurred. Data pipelines interact with various tools and services, leading to potential issues such as data type changes, memory constraints, and buggy code. To address these challenges, Wistia adopted a hub-and-spoke model, where a central data warehouse serves as the single source of truth for business definitions. This approach enables stakeholders to access relevant data in a standardized format, reducing friction and improving data visibility. By leveraging reverse ETL and dbt models, Census helps organizations optimize their pipelines, streamline updates, and reduce overalerting issues, ultimately enabling data teams to focus on more strategic tasks.