How to make Kubernetes rhyme with prod readiness - Tiffany Jernigan and Matthias Haeussler
summary
Kubernetes by itself lacks logging, metrics, ingress, network policy enforcement, persistent-volume plumbing and environment-specific YAML templating, so expect to add tools like kustomize Helm Carvel, delivery with ArgoCD or Flux, and infra tooling such as Cluster API, Velero, Terraform/Pulumi/Crossplane. Monitoring stacks (Prometheus, Grafana, Alertmanager, Jaeger/OpenTelemetry, Fluentd) and security (OPA, cert-manager, SPIRE) are mandatory choices. Managed clusters still need these add‑ons; a built training environment tying them together works for demos but breaks under version drift, so plan for continual maintenance.
AI-written summary · about
chapters (18)
- 0:00Introduction and speakers
- 1:51Why this talk was created
- 3:17What is production readiness
- 5:02Kubernetes basics and features
- 8:54Gaps in vanilla Kubernetes
- 10:21Infrastructure and storage needs
- 11:25Networking and ingress setup
- 12:30Security and observability overview
- 13:53Security policies and access control
- 15:59Workload deployment and tooling
- 17:04CNCF projects for production
- 21:15Helm and templating tools
- 23:02Managed vs unmanaged Kubernetes
- 26:08Training environment example
- 27:10Cloud vs on-prem decision guide
- 28:54When to use managed Kubernetes
- 29:38Key takeaways and recommendations
- 31:45Q&A session
related talks
32mHow to make kubernetes rhyme with prod readiness
29mHow to make Kubernetes rhyme with prod readiness
45mKubernetes Operability Tooling
22mAnatomy of a production grade Kubernetes cluster
43mKubernetes Operability Tooling
39mKubernetes is awesome – but …
51mKubernetes for the Impatient
42mIt (Only) Starts with Containers