Setting Up Feature Flags with Unleash: A Practical Guide

---
apiVersion: apps/v1
kind: Deployment
metadata:
  name: unleash
spec:
  replicas: 2
  template:
    spec:
      containers:
        - name: unleash
          image: unleashorg/unleash-server:5.12
          env:
            - name: DATABASE_URL
              valueFrom:
                secretKeyRef:
                  name: unleash-secrets
                  key: database-url
            - name: SECRET
              valueFrom:
                secretKeyRef:
                  name: unleash-secrets
                  key: admin-secret
            - name: INIT_FRONTEND_API_TOKENS
              value: "frontend:default.frontend-token"
          ports:
            - containerPort: 4242
          readinessProbe:
            httpGet:
              path: /health
              port: 4242
            initialDelaySeconds: 10

PostgreSQL is a mandatory requirement. SQLite is not suitable for production.

Activation Strategies

Unleash divides toggles into several types:

• release — gradual rollout to a percentage of users
• experiment — A/B test with multiple variants
• operational — kill switch for problematic code
• permission — enable for specific roles/groups

For gradual rollout with sticky sessioning (user always sees the same thing):

import { Unleash } from 'unleash-client';

const unleash = new Unleash({
  url: 'https://unleash.internal/api',
  appName: 'api-service',
  customHeaders: {
    Authorization: 'srv-*****',
  },
});

await unleash.start();

// Sticky rollout — user always in the same bucket
const context = {
  userId: user.id,
  sessionId: session.id,
};

if (unleash.isEnabled('new-pricing-page', context)) {
  // show new page
}

The flexibleRollout strategy uses userId or sessionId for consistent hashing — the same user always gets the same result regardless of server.

Unleash Edge for latency reduction

In distributed systems, the SDK makes an HTTP request to the Unleash API on every evaluate. Unleash Edge is a lightweight proxy that caches at the Kubernetes namespace or even pod level:

---
- name: unleash-edge
  image: unleashorg/unleash-edge:latest
  env:
    - name: UPSTREAM_URL
      value: "https://unleash.internal"
    - name: TOKENS
      value: "srv-*****"
  ports:
    - containerPort: 3063
---

The SDK connects to Edge over the cluster network — latency is microseconds instead of milliseconds to the external Unleash server.

Unleash vs LaunchDarkly comparison

Subscription savings when switching from LaunchDarkly are 50-70% for volumes over 10,000 flags, which in monetary terms means reducing costs from $2,000/month to $600-1,000/month. However, Unleash requires engineering hours for maintenance — on average 4 hours per month, at a rate of $100/hour adds $400/month in additional costs. Net savings can reach $1,000-1,600/month.

How we set up Unleash: step-by-step plan

1. Infrastructure audit — assess current microservices, number of flags, latency requirements.
2. Deploy server and database — set up Unleash + PostgreSQL in Kubernetes using Helm or Kustomize.
3. Integrate SDK — connect clients (Node.js, Go, Python) to the server or Edge, configure tokens.
4. Configure strategies — create custom strategies for canary, A/B, kill switch, import existing toggles.
5. Monitoring and alerts — set up Grafana dashboards, alerts on SLO and MTTR, write runbook for incidents.

Why choose Unleash over LaunchDarkly?

Unleash is better than LaunchDarkly in scenarios where data control and cost are important. For volumes over 10,000 flags, subscription savings are 50-70%, but you pay with Ops time. Unleash is fully open-source — you can fork and customize. LaunchDarkly gives less infrastructure hassle but ties you to the cloud.

How to reduce flag evaluation latency?

Use Unleash Edge — a lightweight proxy that caches flags in memory. Edge can be placed in each Kubernetes namespace so SDKs connect locally. At 1,000,000 evaluates per hour, Edge reduces server load by 10x and guarantees latency <1ms.

What's included in turnkey Unleash setup

How we ensure deployment reliability?

We use GitOps with ArgoCD, cover configuration with tests, and provide runbooks for incidents. Experience from 30+ projects helps avoid typical mistakes in RBAC, NetworkPolicy, and scaling. We guarantee stable operation: SLO 99.9% for the Unleash server.

Implementation timeline and how to order

Server deployment + basic SDK integration — from 1 day. Strategy configuration, Edge proxy, import of existing flags — from 1 to 2 days. Cost is calculated individually after an infrastructure audit.

Contact us for an audit of your project. Get a consultation on the optimal Unleash configuration. Order feature flag setup to speed up your release cycle. Write to us, we are always available.