Event-based Continuous Delivery: The End of the Monolithic Pipeline
summary
Argues that replacing monolithic pipelines with an event-driven control plane decouples process from tools and shortens maintenance. Implements a CNCF sandbox called kptn/captain running on Kubernetes, using Knative/CloudEvents, Helm, JMeter, Dynatrace (or Prometheus), GitLab CI hooks, and chaos tools (Litmus/Gremlin). Uses a shipyard file for stages, SLO-based quality gates and automated promotion; demo shows build 4 scoring 100 in staging but failing in prod and being automatically rolled back, proving the workflow and remediation loop work.
AI-written summary · about
chapters (9)
- 0:00Introduction and speaker background
- 3:07Problems with traditional pipelines
- 5:15Event-driven delivery solution
- 9:27Demo of event-driven workflow
- 13:20Automating approvals with SLIs and SLOs
- 16:27Demo of quality gates
- 18:52Event-driven SRE for remediation
- 20:59Summary of Captain's benefits
- 21:40Q&A session begins
related talks
52mThe Future of Continuous Delivery
47mModern continous delivery on Azure Kubernetes Service
41mThe Pipeline-Driven Organization
26mWhere Are We Going? The Future of Continuous Delivery
36mA Guide to Event-driven SRE-inspired DevOps
41mAll Things Continuous
30mModern Continuous Delivery
17mDriving Continuous Delivery: Trunk-Based Deployment Strategies with PipeCD in Large Scale