Our Journey in Integrating GitOps Tools Into Delivery Pipelines
summary
Built an integration that wires Argo/Flux notifications to a controller that runs k6 load testing and Prometheus SLI evaluations, creates auto-promote PRs and auto-merges when quality gates pass; demo uses three services (Potato Head app) and trace testing via OpenTelemetry. Rollbacks prove brittle and a “ping‑pong” between tools emerges at scale; current usage is small scale only. To fix that, a Kubernetes operator called Captain Lifecycle Controller annotates workloads, runs pre/post tasks and evaluations, and feeds deployment decisions back into GitOps.
AI-written summary · about
chapters (10)
- 0:00GitOps integration journey
- 2:09What is Captain
- 4:19Imperative Captain challenges
- 6:24GitOps workflow with Captain/Argo/Flux
- 10:42Demo: automated testing & promotion
- 14:35Current approach challenges
- 19:51Context propagation limits
- 21:36Captain Lifecycle Controller
- 25:36Findings & future of GitOps
- 26:41Q&A
related talks
35mPanel: GitOps User Stories
07mNavigating Multi-stage Deployment Pipelines via the GitOps Approach
07mIntroduction to GitOps
16mWant to extend GitOps to Infrastructure and Workloads? Here's how.
29mCombining Progressive Delivery with GitOps and Continuous Delivery
29mObservability Driven Automation: Beyond GitOps with Keptn
31mLeveraging GitOps to wrangle your clusters and projects
45mFrom DevOps to GitOps: Supercharging Your Kubernetes Workload Deployments