5 Years with K8s: Container Deployment at Pôle Emploi
summary
A five‑year on‑prem Kubernetes platform at Pôle Emploi running dozens of clusters (TKGI/Tanzu and RKE2 managed by Rancher) on VMware, with automated cluster pipelines using Concourse and GitHub. Observability is implemented with Prometheus, Grafana and Thanos; backups with Velero; tracing with OpenTelemetry and AppDynamics; and SRE practices use SLOs, autoscaling and fleet management to maintain consistent versions and automated rollouts. Adoption grew from 2019 to over 20,000 pods by 2023, and right‑sizing reduced resource usage about fourfold while NPS reached 75.
AI-written summary · about
chapters (17)
- 0:00Introduction
- 1:55Container adoption
- 3:42Infrastructure pre-K8s
- 5:27TMJ project goals
- 7:37First app live
- 9:26Observability improvements
- 11:56SLO and SRE dashboards
- 15:12Hybrid clusters RK2
- 17:25Five years of K8s
- 18:08User NPS feedback
- 19:13Security approach
- 21:23Platform team
- 22:49Success factors
- 25:16Ecosystem tools
- 27:43UX simplification
- 29:06Service automation
- 30:07Team best practices
related talks
35mKubernetes journey
40mFrom dev to prod. One Year of Using Kubernetes in Production
17mOur Journey to Kubernetes at Scale
53mImplementing Kubernetes on with confidence
29m11 5 years of kubernetes and looking to the future
29mA history of GOV.UK's infrastructure: the long voyage to Kubernetes
46mAround & After Kubernetes: The Principles and Ideas that Guide Us
35mThriving With Kubernetes On-Call: Best Practices & Lessons Learned