Data practitioners can apply support data, including from agent and in-app interactions, to streamline and automate support processes and improve the product experience. In this post, Nimesh Agarwal, Aravid Ranganathan, and Pallavi Nagesharao describe how Uber collects and analyzes customer issues through its Support Insights Platform. The platform, which processes, stores, and serves metrics, leverages Kafka for messaging, Flink for stream processing, Spark for batch processing, and Pinot as its database. To support disparate use cases (including several dashboards and web applications), Uber also developed a Golang microservice that serves data from Pinot and other production databases and performs optimizations to improve data quality. In addition, Uber developed its own ReactJS-based visualization tool for workforce management to provide near-real-time data freshness and expose metric computation logic.