Kubernetes Platform Comparative
Jump to navigation
Jump to search
Kubernetes Distributions (Prominent)
- Kubeadm - https://github.com/kubernetes/kubeadm
- AKS - https://azure.microsoft.com/en-us/services/kubernetes-service/
- Microk8s - https://microk8s.io/
- OKD - https://github.com/openshift - (Fork or dev repo of OpenShift) - Selected commits merged to upstream to Openshift on approved PRs
- OpenShift - https://github.com/openshift - IBM/Redhat Commercial -
- VMWWare (7) - https://tanzu.vmware.com/kubernetes-grid - https://bitnami.com/ - https://www.vmware.com/content/dam/digitalmarketing/vmware/en/pdf/vsphere/vmw-vsphere7-solution-brochure.pdf
Terminologies, Definitions & Clarifications
- k8s vs k3s - https://www.civo.com/blog/k8s-vs-k3s
More Comparative Grids
Possible Open Operators
- You can always go alone and create your own or you can synergize by working with an existing one like ...
- https://github.com/CrunchyData/postgres-operator
- https://gitlab.com/frenchtoasters/zabbix-operator
Minimal Kubernetes Architecture - 5 VMs at a minimum for complete HA redundancy
- https://kubernetes.io/docs/concepts/overview/components/
- 3 vms running Kubernetes Controllers/Nodes
- 2 vms running adc/layer7 load balancing - This could be as simple as VRRP(keepalived & nginx or hapeproxy) There are commercial offerings of this as well.
- vms live on different baremetal vm hosts so they aren't a point of failure.
Thoughts on platforms and Operators
- All platforms should be portable
- All applications should be migratable.
- You should be allowed to use best of bread and not locked in
- Decisions should be logical and not emotional. Choices need to be logically backed. You and others will feel pain if they aren't