Service Delivery Index for Reliability

This thread tracks the evolution of Slack's Service Delivery Index (SDI) as a key metric for measuring and driving service reliability. Initial posts detailed the concept of SDI-R as a composite metric of API availability and user workflow success, its importance in transitioning from reactive 'hero engineering' to proactive reliability engineering, and its role in setting customer expectations and aligning teams. This post details the program's adoption, focusing on the visibility and accountability mechanisms, including leadership reporting and linking regressions to specific teams. It also covers scaling strategies through tools like the Webapp Ownership Tool and Omni, and organizational forums like the Engineering Monday Meeting. Lessons learned include the need to weight API requests by importance, the value of service-specific alerting for long-term maintainability, and the continuous evolution of metrics and processes to adapt to new products and infrastructure.