How would you take a service from 99.9% to 99.99% availability?

Summary

First confirm the business needs it, because it cuts allowed downtime from about 43 minutes to about 4 minutes a month. Then analyse where the current budget goes, and attack the largest causes: remove single points of failure (multi-zone, redundancy), make changes safer (canaries, automated rollback), speed up detection and mitigation (burn-rate alerts, automated failover), and degrade gracefully when dependencies fail.

Longer Answer

At four nines, humans are too slow to be the recovery mechanism: an engineer being paged, waking up and logging in already uses the monthly budget. So recovery must be automatic (health-check-based failover, automated rollback, load shedding), and changes must be staged so a bad one only affects a small slice. Check critical dependencies: you cannot exceed their availability without redundancy or fallbacks. Measure progress against the SLO, and price the cost (extra regions, engineering time) explicitly.

Key Points

• Confirm the need; 4 minutes/month.
• Spend analysis: fix the biggest budget consumers.
• Automated recovery; humans are too slow.
• Dependencies cap you.

What interviewers look for

Budget-driven prioritisation and the insight that manual response cannot meet four nines.