🖥️ System Testing:

System testing verifies the entire integrated system meets requirements, testing it as a complete product.

System Testing = Integration Testing + Full Environment

Not just: Do services talk to each other? But: Does the entire system work in production-like environment?

Real-World Analogy:

Like testing an entire airplane before first passenger flight:

Component Testing: Each part tested (✅ done)

Integration Testing: Parts work together (✅ done)

System Testing: Complete airplane test

✓ Can it taxi on runway?

✓ Can it take off?

✓ Can it fly at cruising altitude?

✓ Can it land safely?

✓ Do all systems work together (navigation, fuel, hydraulics)?

✓ Can it handle turbulence?

Test the COMPLETE system as passengers will experience it!

System Test Example: Banking Application

System Under Test:

• Web frontend (React)

• Mobile app (iOS/Android)

• API gateway

• Authentication service

• Account service

• Transaction service

• Notification service

• Database cluster

• Message queue

• External payment gateway

System Test Scenario: "Fund Transfer"

Prerequisites:

✓ Production-like environment deployed

✓ Test data loaded (accounts, balances)

✓ All services running

✓ External systems configured (mocked or sandbox)

Test Steps:

1. User logs in via web app

   Verify: ✓ Session created

   Verify: ✓ Dashboard loads with account balance

2. User navigates to transfer page

   Verify: ✓ Account list populated

   Verify: ✓ Available balance correct

3. User initiates transfer ($500)

   Input: Recipient account, amount, note

   Verify: ✓ Validation works (sufficient funds)

   Verify: ✓ Confirmation screen shown

4. User confirms transfer

   Verify: ✓ Transaction processes

   Verify: ✓ Sender balance decreases

   Verify: ✓ Recipient balance increases

   Verify: ✓ Transaction logged in database

5. System sends notifications

   Verify: ✓ SMS sent to sender

   Verify: ✓ Email sent to sender

   Verify: ✓ Push notification to mobile app

6. User checks mobile app

   Verify: ✓ Transaction appears in history

   Verify: ✓ Balance updated

   Verify: ✓ Transaction details correct

7. System processes batch jobs (overnight)

   Verify: ✓ Daily summary generated

   Verify: ✓ Audit logs created

   Verify: ✓ Compliance reports generated

All components working together in complete system! ✅

Types of System Testing:

1. Functional System Testing

Test: Complete business workflows

• User registration flow

• Purchase flow

• Refund flow

• Account closure flow

Each flow touches multiple subsystems

2. Non-Functional System Testing

Test: System qualities

• Performance (response times under load)

• Security (penetration testing)

• Usability (can users navigate easily?)

• Reliability (uptime, error rates)

• Scalability (handles growth?)

3. Recovery Testing

Test: System recovery from failures

Simulate failures:

1. Kill database mid-transaction

   Expected: ✓ Transaction rolled back

   Expected: ✓ User sees error message

   Expected: ✓ System remains stable

2. Restart application server

   Expected: ✓ In-flight requests handled gracefully

   Expected: ✓ New requests succeed after restart

3. Network partition

   Expected: ✓ System detects partition

   Expected: ✓ Failover to backup

System Testing Environment:

Production-like "Staging" Environment:

[Load Balancer] (Same config as prod)

↓

[App Servers] (Same scale as prod)

↓

[Databases] (Replicated setup)

↓

[Cache Layer] (Redis cluster)

↓

[Message Queue] (RabbitMQ/Kafka)

↓

[External APIs] (Sandbox/mocked)

As close to production as possible!

Key characteristics:

• 🎭 End-to-end scope: Entire system, all components
• 🌍 Production-like environment: Realistic infrastructure
• 📋 Requirements verification: Does it meet all specs?
• 🔄 Multiple user workflows: Real usage scenarios

Key Takeaways

1. System testing validates the entire integrated system end-to-end — testing the full stack as a user would experience it
2. System tests run in a production-like environment — same infrastructure, configuration, and data characteristics
3. These tests are slow and expensive — run them less frequently than unit or integration tests (nightly, not on every commit)
4. System tests catch environment-specific issues — configuration errors, network policies, and resource limits that unit tests miss