High‑Availability Architecture and Reliability Practices from a Former Google SRE

1. Two Major Factors Determining Availability

The author emphasizes that availability is fundamentally about Service Level Agreements (SLA) and introduces the concepts of Mean Time Between Failures (MTBF) and Mean Time To Recover (MTTR) as the two quantitative factors that drive a service’s uptime.

2. High‑Availability Solutions

Practical approaches include increasing redundancy (e.g., deploying N+2 instances instead of N+1), ensuring instances are independent and equally capable, and implementing sophisticated traffic‑control mechanisms that can isolate, quarantine, or block problematic requests based on business‑level attributes.

Change‑management techniques are highlighted: thorough offline testing, staged gray‑release deployments with carefully selected canary users, and mandatory rollback support built into the release process.

3. Seven‑Level Availability Maturity Model

The article presents a tiered chart ranging from Level 1 (crash with data loss) to Level 7 (failover with negligible user impact), describing the technical and operational investments required to progress through each level.

Q & A

Answers cover topics such as monitoring and SLA measurement, low‑level OS and hardware caching concerns for “crash without data loss,” the importance of error budgets, tools used for managing millions of servers, and concrete strategies for handling problematic traffic and safe rollbacks.