Kubernetes Core Concepts

K8s core architecture is split into focused topics, each with Python analogies for quick intuition.

Navigation

Topic	Content
Ingress & Service	How requests flow from outside into Pods, 3 Service types, internal DNS
Workload Types	Deployment / StatefulSet / DaemonSet / Job / CronJob — which to pick
Storage	StorageClass / PVC / PV three-layer model, gp2 vs gp3, IOPS vs Throughput
CronJob	Scheduled task config, concurrencyPolicy, typical use cases
Observability	Pod log queries, multi-replica log problem, EKS vs Lambda selection

---

Quick Mental Model

Incoming traffic:
  Internet → Ingress (routing rules) → Service (finds Pods) → Pod (runs your code)

Workload types for an inference system:
  - Model server (long-running)  → Deployment
  - Database / vector DB         → StatefulSet
  - Log collector                → DaemonSet
  - Batch inference              → CronJob
  - DB migration                 → Job

# The whole K8s cluster is like a deployment platform for a Python app:

# Deployment  = uvicorn running FastAPI server (multiple replicas)
# StatefulSet = PostgreSQL (data must persist, name must be stable)
# DaemonSet   = dcgm-exporter on every GPU server (node-level agent)
# Job         = python migrate.py (runs once and exits)
# CronJob     = crontab's K8s equivalent (triggers Jobs on schedule)

---

Quick Reference

For detailed K8s glossary see K8s Glossary.