Scaling Flux at BT: A Multi-Cluster Telco Migration Journey
summary
BT migrates a mature Terraform + GitLab deployment into a GitOps pipeline using Flux across about 30 clusters, replacing per-cluster state files and improving self-healing via reconciliation loops. They convert imperative scripts into K8s Jobs, confront overlay sprawl, stalled Helm releases and long dependency chains, and build ephemeral dev environments for NetBox and SDN tooling. Current progress is ~30% with a goal of ~80% by quarter end; upcoming work includes Flagger/Flux Operator for canaries, auto image updates, and safer DB migrations.
AI-written summary · about
chapters (8)
- 0:00Flux migration intro at BT
- 2:48Why Flux at BT
- 5:38Migration challenges and lessons
- 8:08Scaling Flux adoption
- 12:43Netbox to Flux migration
- 15:35Testing with Flux pipelines
- 16:16Current state and future
- 20:59Q&A
related talks
07mFluxCD - Gitops For All Sizes
06mProject Lightning Talk: Flux - What is it & What's New?
39mEnhance Your Kubernetes Experience with GitOps & Flux
42mIntro: Flux
01h 01mManaging Kubernetes the GitOps way with Flux
27mOrchestrating Multi-Tenancy Kubernetes Environments with Flux
36mCD and GitOps with flux @AdoreMe
10mCNCF Project Update: Flux
