Exercises › Multi-Region Deployment Language › Failover Vocabulary

Failover Vocabulary

5 questions · Multi-Region Deployment Language

1. A team states: "We have an RPO of 15 minutes." What does this mean in plain English?

The system will recover and be available again within 15 minutes of a failure In the worst case, we can lose up to 15 minutes of data if a failure occurs Failover must be completed in under 15 minutes to meet our SLA

2. An engineer explains: "Our RTO is 4 hours — that means if the primary region goes down at midnight, we must have the secondary fully serving traffic by 4 AM." What does RTO measure?

The maximum acceptable amount of data that can be lost during a failure The maximum acceptable time for the system to recover and resume normal operations after a failure The time between scheduled maintenance windows for each region

3. The team discusses automatic failover vs. manual failover. Which statement correctly describes the trade-off?

Automatic failover is always preferable — manual failover is only used in legacy systems Automatic failover is faster (no human delay) but carries risk of triggering on false positives; manual failover requires human decision-making but avoids unnecessary failovers Manual failover is faster because a human can assess the situation faster than automated systems

4. After recovering from an incident, the team plans to failback. What does this operation involve?

Running a post-incident review to document what went wrong during the failover Returning traffic and operations to the original primary region after it has been restored Adding extra capacity to the secondary region to prevent future failures

5. A SRE warns: "We have significant replication lag between regions right now." Why is this relevant to failover planning?

High replication lag means failover will be slower to initiate High replication lag means if we fail over now, we will lose more data than our RPO allows — the secondary is behind the primary Replication lag only affects read performance, not failover data integrity

Exercise complete!

out of 5 questions

← Back to Multi-Region Deployment Language All exercises