What if your deployment pipeline could sense failure and heal itself before users notice?

In this session, we show how to build self-healing progressive delivery pipelines using Argo Rollouts powered by real-time observability signals. Integrating Prometheus, OpenTelemetry, and Linkerd, we transform rollouts from time-based experiments into signal-driven systems that auto-react to performance degradation.

See how:
- A canary detects rising latency and auto-rolls back early before wider exposure
- Error spikes trigger KEDA-based scaling and pause Rollouts until stable
- Service-mesh insights feed AnalysisTemplates for closed-loop remediation

While we use specific tools, the patterns are universal: any mesh, any metrics stack, any automation framework. You’ll take home templates, blueprints, and decision frameworks that make Argo Rollouts the control plane for resilience, ushering in the next phase of progressive delivery: Automated Resilience Engineering.