Kubernetes Fundamentals: Objects, YAML, Labels, Namespaces, and Workloads

Kubernetes Fundamentals: Objects, YAML, Labels, Namespaces, and Workloads

Kubernetes is easiest to understand as an API of desired-state objects. You declare what should exist, and controllers continuously work to make actual state match it.

flowchart LR
    Y[YAML / kubectl] --> A[Kubernetes API]
    A --> C[Controllers]
    C --> W[Workloads / Services / Storage]
    W --> S[Status]

Kubernetes Objects

Most resources have:

• spec — what you want.
• status — what Kubernetes currently observes.

A Deployment spec might request three replicas; status reports how many are actually ready.

Declarative YAML

A basic object includes:

apiVersion: apps/v1
kind: Deployment
metadata:
  name: api
spec:
  replicas: 3

Prefer version-controlled declarative manifests or an equivalent configuration-as-code workflow over undocumented manual cluster changes.

Labels and Selectors

Labels organize resources and connect controllers/services to the objects they manage.

metadata:
  labels:
    app: orders
    environment: production

Selectors must be designed carefully. Overlapping controller selectors can cause conflicting ownership.

Use annotations for non-identifying metadata that does not need selection.

Namespaces

Namespaces scope names and help organize teams/projects and apply policies or quotas.

They are not automatically a security boundary. RBAC, NetworkPolicy, quotas, and other controls still matter.

For production, avoid dumping everything into the default namespace.

Core Resource Map

Pod        → running container group
Deployment → manages stateless Pods
Service    → stable network identity
ConfigMap  → non-secret configuration
Secret     → sensitive configuration object
PVC/PV     → persistent storage
Job        → run-to-completion work
Namespace  → resource scope

Learn these before jumping into the wider Kubernetes ecosystem.

kubectl Is an API Client

Commands such as:

kubectl get pods
kubectl describe deployment api
kubectl apply -f deployment.yaml

are simply ways to interact with the Kubernetes API. The API—not kubectl—is the real platform boundary.

Production Habits From Day One

• Use consistent recommended labels such as app.kubernetes.io/* where useful.
• Keep manifests/configuration version-controlled.
• Set resource requests for workloads.
• Use readiness probes for traffic eligibility.
• Separate sensitive Secrets from normal configuration.
• Use namespaces/policies intentionally rather than creating dozens without purpose.

Final Takeaway

The Kubernetes foundation is small: objects, desired state, controllers, labels/selectors, namespaces, and a handful of workload/network/storage resources. Understand that object model first; most advanced Kubernetes features are extensions of the same reconciliation idea.

References

• Objects in Kubernetes
• Labels and Selectors
• Namespaces
• Kubernetes Concepts