Progressive Delivery with Observability
summary
A concrete recipe for progressive delivery that ties runtime telemetry to CI-driven promotion and rollback: a JavaScript/Express catalog app emits OpenTelemetry traces and versioned service tags, CircleCI queries Honeycomb for error counts and user-scoped latency, and a scripted threshold (e.g., X errors per minute for released cohort) fails the pipeline to stop promotion. Live traces show which user IDs and service spans fail, letting teams programmatically canary, promote, or roll back based on real production signals rather than dashboards or guesses.
AI-written summary · about
chapters (7)
- 0:00Introduction and background
- 1:04What is observability
- 3:55Improving deployments with observability
- 7:09What observability looks like
- 9:19Demo setup and problem statement
- 14:40Testing in production with tracing
- 15:43Growing with application changes
related talks
14mProgressive Delivery with Observability
22mProgressive Delivery: The Logical Evolution CI/CD
36mProgressive Delivery in the Kubernetes era
51mProgressive Delivery in the Kubernetes era
38mProgressive Delivery: Patterns & Benefits of Decoupling Deploy from Release
35mProgressive Delivery with Argo Rollouts
43mCloud-Native Progressive Delivery
28mEnd-to-end CI/CD observability with open source