What is the difference between fault tolerance and high availability?

System design interviewers love to ask:

“What’s the difference between fault tolerance and high availability?”

At first glance, they sound the same. But they describe two levels of resilience in distributed systems — one prevents failure from happening, the other minimizes its impact.

1️⃣ Quick definition to use in interviews

• Fault tolerance → The system continues operating even if a component fails.
• High availability (HA) → The system recovers quickly when a failure occurs.

In short:

Fault tolerance = No downtime at all. High availability = Minimal downtime.

2️⃣ Example that interviewers understand easily

Imagine you’re running a flight booking system:

• If one server dies and users never notice → that’s fault tolerance.
• If one server dies, users see a 5-second retry before it switches over → that’s high availability.

Both designs ensure reliability, but the tolerance approach costs more and is used where downtime is unacceptable (e.g., healthcare, finance).

3️⃣ How they work under the hood

🔗 Read: High Availability System Design Basics

4️⃣ How to design for both in interviews

If asked “How would you make this system fault tolerant?”, mention:

• Replication across zones (multi-AZ or multi-region)
• Active-active architecture
• Stateless app servers behind a load balancer
• Automatic failover for databases
• Health checks and retry logic

And follow up with:

“If 100% fault tolerance isn’t cost-effective, I’d aim for 99.99% availability with fast failover.”

This shows pragmatic engineering judgment — something interviewers value deeply.

🔗 Related: Data Replication Strategies in System Design

5️⃣ Example phrase to use during interviews

When comparing systems:

“Our chat app can tolerate a single server failure without user impact — that’s fault tolerance. But if an entire region goes down and the system recovers in seconds, that’s high availability.”

This makes your answer crisp, memorable, and technically correct.

💡 Interview Tip

End your answer with something like:

“I’d combine both — build for high availability first, then make critical components fault tolerant.”

This demonstrates layered thinking and real-world prioritization.

🎓 Learn More

Strengthen your reliability and failover skills inside 👉 Grokking the System Design Interview and Grokking System Design Fundamentals. Both include diagrams and examples showing how companies like Netflix and Amazon achieve fault tolerance.