On-Premises Kubernetes

Run Kubernetes where the cloud can’t go.

Not every workload belongs in the cloud. Data sovereignty, latency requirements, regulatory constraints, and economics drive enterprises to run Kubernetes on their own hardware. This track covers everything from datacenter planning to day-2 operations — the knowledge that most free resources skip because it’s not glamorous, but it’s where a massive share of production Kubernetes actually runs.

---

Learning Path

Planning & Economics (5 modules)
        │
        ▼
Bare Metal Provisioning (4 modules)
        │
        ├── Networking (6 modules)
        ├── Storage (5 modules)
        └── Multi-Cluster & Platform (5 modules)
                │
                ▼
        Security & Compliance (8 modules)
                │
                ▼
        Day-2 Operations (9 modules)
                │
                ▼
        Resilience & Migration (3 modules)
                │
                ▼
        AI/ML Infrastructure (6 modules)

Sections

Section	Modules	Focus
Planning & Economics	5	Server sizing, cluster topology, TCO, cloud vs on-prem, FinOps & chargeback
Bare Metal Provisioning	4	PXE, MAAS, Talos, Sidero/Metal3
Networking	6	Spine-leaf, BGP, MetalLB, DNS/certs, cross-cluster, service mesh
Storage	5	Ceph/Rook, local storage, object storage (MinIO), database operators
Multi-Cluster & Platform	5	vSphere/OpenStack, vCluster/Kamaji, Cluster API, fleet management, active-active
Security & Compliance	8	Air-gapped, HSM/TPM, AD/LDAP, SPIFFE, Vault, policy-as-code, zero-trust
Day-2 Operations	9	Upgrades, firmware, observability, capacity, self-hosted CI/CD & registry, serverless
Resilience & Migration	3	Multi-site DR, hybrid connectivity, cloud repatriation
AI/ML Infrastructure	6	GPU nodes, private training, LLM serving, MLOps, AIOps, HPC storage

51 modules total (30 existing + 21 new from #197). From “should we go on-prem?” to “how do we train LLMs on our own GPUs.”

---

Readiness Check

This is an advanced track. You should already be comfortable with:

• core Kubernetes concepts and troubleshooting
• Linux networking, storage, and security basics
• day-2 operational thinking such as upgrades, observability, and failure handling

If you are not there yet, strengthen those areas first through Prerequisites, Linux, Kubernetes Certifications, and optionally Platform Engineering.

Safest Route Into On-Prem

Prerequisites
   |
Linux
   |
CKA-level Kubernetes understanding
   |
Platform / SRE thinking
   |
On-Premises

You do not need to finish every module in those tracks first, but you do need the operational maturity they represent.

Red Flags That You Are Entering Too Early

• you still struggle with Linux networking, storage, or service management basics
• you have not yet operated Kubernetes under failure, upgrade, or capacity pressure
• you mainly want a simpler first cluster rather than private-platform design

If those are true, stay in Prerequisites, Linux, or Kubernetes Certifications a bit longer.

Common Bridge Routes Into On-Prem

Coming from	Safest bridge	What to prove before going deeper
Kubernetes Certifications	CKA -> Linux depth -> On-Prem Planning/Provisioning	kubeadm, networking, storage, and troubleshooting are not fragile for you
Cloud	managed Kubernetes -> Architecture Patterns -> On-Prem Planning	you can reason about tradeoffs without assuming the cloud will always provide the control plane around you
Platform Engineering	SRE / GitOps / Networking -> On-Prem Operations	you already think in day-2 systems, not only cluster setup
AI/ML Engineering	local-first AI -> AI Infrastructure -> On-Prem AI/ML Infrastructure	you understand the jump from a workstation or notebook workflow to private GPU fleet operations

---

Prerequisites

• Fundamentals — Cloud Native 101, K8s Basics
• Linux — networking, storage, security hardening (includes LFCS)
• Certifications — CKA (cluster architecture, kubeadm) is required
• Recommended: CKS for security modules
• Recommended: Platform Engineering for SRE and operations modules

Who This Is For

• Infrastructure engineers building private Kubernetes platforms
• Platform teams evaluating on-prem vs cloud for their organization
• SREs operating bare metal or private cloud Kubernetes clusters
• Architects designing multi-site, air-gapped, or hybrid environments
• Budget owners calculating TCO and making build-vs-buy decisions

Common Next Steps

• go to Platform Engineering if you want deeper reliability, delivery, and platform-discipline context around private infrastructure
• go to AI/ML Engineering if your on-prem goal includes private model training or local-first AI systems
• go back to Cloud if you need a cleaner contrast between managed-cloud assumptions and private-infrastructure tradeoffs before committing to this path