Quality Gates in Production - How We Turn OpenTelemetry Signals into Deployment Decisions

Learn how to build automated quality gates that make deployment decisions based on production monitoring data in this 21-minute conference talk from SREcon25 EMEA. Discover a practical implementation using open-source tools including GitHub Actions, ArgoCD, TestContainers, Playwright, and OpenTelemetry to validate deployments before they reach users. Explore how to correlate deployment events with error rates, latency, and business metrics using OpenTelemetry traces, logs, and metrics, while examining quality gate criteria that catch regressions across data pipelines and application services. See the actual pipeline in action and understand how this approach proves particularly valuable for data-heavy and AI workloads where traditional health checks provide limited insight. Gain insights into implementing quality gates for AI services using MLFlow for experiment tracking and model management, including validation of model accuracy, drift detection, and inference performance verification. Take away practical quality gate patterns you can implement with existing tools, specific metrics that indicate deployment success, and lessons learned from operating this system in production to improve deployment confidence while maintaining development velocity.