Skip to content

Platform Component Examples

Additional examples of the component replacement pattern for infrastructure migrations.

See Platform Component Replacement for the core pattern and database migration example.

Pattern Consistency

All infrastructure component replacements follow the same build-replace-remove workflow. The examples below use different technologies but the same underlying pattern.

Service Mesh Replacement

Old: Linkerd New: Istio

# Phase 1: Deploy Istio alongside Linkerd
# Phase 2: Migrate namespaces one at a time
# Phase 3: Remove Linkerd annotations
# Phase 4: Uninstall Linkerd

# No application changes
# Service-to-service communication continues
# mTLS preserved throughout

Kubernetes Operator Upgrade

Old: Operator v1 (CRD v1alpha1) New: Operator v2 (CRD v1)

# Phase 1: Deploy Operator v2 with CRD v1
# Phase 2: Operator v2 converts v1alpha1 → v1
# Phase 3: Existing resources still work (conversion webhook)
# Phase 4: Update manifests to v1 over time
# Phase 5: Remove Operator v1

# Zero downtime for workloads
# CRD conversion handles compatibility

Storage Backend Migration

Old: AWS EBS volumes New: AWS EFS shared filesystem

# Phase 1: Create EFS filesystem
# Phase 2: Mount EFS alongside EBS
# Phase 3: Sync data from EBS to EFS
# Phase 4: Update PVC to point to EFS
# Phase 5: Verify application reads from EFS
# Phase 6: Remove EBS volumes

# Application sees same mount point
# Data copied in background

Comments