Fault Tolerance and Redundant Systems Online Quiz Questions

A service continues returning correct results, but its latency is so high that it violates the application's timing requirement. Which term best describes this condition?

A service has three application servers, but all of them depend on one unreplicated database. What is the unreplicated database in this design?

An application process is running, but it cannot currently serve requests because a required dependency is unavailable. Which health check should normally prevent the instance from receiving traffic without necessarily restarting it?

Which two statements correctly describe synchronous replication? Select all that apply.

Which two patterns directly help contain the blast radius of a failing dependency or overloaded tenant? Select all that apply.

True or false: Replication alone is sufficient protection against accidental deletion or corruption because every replica contains a copy of the data.

True or false: When one of two servers fails, failover automatically gives the surviving server enough capacity to handle all the original traffic safely.

What is the standard term for the maximum acceptable amount of data loss after a disaster?

What failure-isolation pattern temporarily stops requests to a dependency that is repeatedly failing?

A consensus cluster has five nodes, and two nodes fail. Explain why the cluster can generally continue making progress, and state the key assumption that makes this possible.

After repairing and resynchronizing a failed replica, which next action best reduces the risk of returning an unsafe system to full production load?

A service process is still running, but it cannot currently serve requests because a required dependency is unavailable. Which health check should normally prevent this instance from receiving new traffic without necessarily restarting it?

A database must remain available during an ordinary node failure and also recover from accidental deletion that is replicated to every live copy. Which design best addresses both requirements?

True or false: A performance problem can never become a system failure if the system eventually returns correct results.

A disaster-recovery requirement says that service must be restored within 15 minutes and that no more than approximately 5 minutes of recently committed data may be lost. What is the RTO, expressed in minutes?

What is the exact health-check term for a check that determines whether an instance should receive traffic?

Two independent components are both required for a request to succeed. Each has availability 0.99. What is the approximate combined availability of the series system?

An API repeatedly calls a dependency that is timing out. The system should temporarily stop sending calls to that dependency and resume them only after it appears healthy. Which failure-isolation pattern is most appropriate?