Q: Engineering squads use three different paid SaaS feature flagging tools, leading to high license costs, scattered audit logs, and inconsistent SDK implementations. How do you standardize feature flagging across the enterprise using OpenFeature and self-hosted Unleash on Kubernetes?

Deploy Unleash with PostgreSQL backend and horizontal pod autoscaling. Configure enterprise OIDC authentication and audit logging.

apiVersion: apps/v1
kind: Deployment
metadata:
  name: unleash-server
  namespace: platform-flags
spec:
  replicas: 3
  template:
    spec:
      containers:
        - name: unleash
          image: unleashorg/unleash-server:v5.5
          env:
            - name: DATABASE_URL
              valueFrom: { secretKeyRef: { name: unleash-db, key: url } }

Incorporate OpenFeature SDKs into application templates. Applications write code against the neutral OpenFeature API rather than proprietary vendor SDKs.

// Go application using OpenFeature
openfeature.SetProvider(unleash.NewProvider(unleash.WithUrl("https://flags.acme.internal")))
client := openfeature.NewClient("checkout-service")
isNewCheckoutEnabled, _ := client.BooleanValue(ctx, "checkout-v2-enabled", false, evalCtx)

Build an automated audit job that queries Unleash for flags older than 90 days that have stayed at 100% rollout. Emit Backstage alerts prompting squads to clean up dead code branches.