cloudnative wiki

Non-Functional Requirements & Cluster Operations

1 min read

Non-Functional Requirements & Production Readiness

Operational excellence guides addressing cluster architecture, reliability, governance, and capacity planning.

Pillar Directory

1. Reliability & Availability

  • High Availability: Control plane topologies, etcd quorum, multi-AZ nodes, and Pod Disruption Budgets (PDBs).
  • Disaster Recovery: RTO/RPO targets, multi-region failover, and cluster rebuild workflows.
  • Backup & Restore: Velero backups, etcd snapshots, volume snapshots, and verification drills.
  • Chaos Engineering: Validating cluster resilience using Chaos Mesh and LitmusChaos.

2. Scalability & Performance

  • Autoscaling Strategy: Combining HPA (scale-to-zero), VPA, KEDA, and node provisioners (Karpenter / Cluster Autoscaler).
  • Performance Tuning: Kernel limits, cgroup v2, kubelet parameters, and runtime configurations.
  • Cost Optimization: Workload rightsizing, Spot instance management, and bin-packing.

3. Security & Governance

  • Security Baseline: Pod Security Standards, NetworkPolicy default-deny, and policy engines (Kyverno / OPA Gatekeeper).
  • Multi-Tenancy: Namespace boundaries, vCluster, quota enforcement, and network isolation.
  • OIDC Authentication: Connecting enterprise IdPs (Okta, Keycloak, Dex) to the Kubernetes API server.

4. Lifecycle & Upgrades

  • Upgrade Strategy: Zero-downtime control-plane and worker-node upgrades for self-managed and cloud clusters.
  • API Deprecations: Detecting removed APIs, version gates, and migration schedules.

12 items under this folder.