Building a system is half the job. The other half is keeping it up: knowing when it is unhealthy, waking the right person for the right reason, fixing an outage fast, and changing and scaling it without causing the next one. This course covers site reliability engineering, one operational decision at a time.

Each section takes one question that SRE and production-engineering interviews actually ask and answers it as a standalone topic, with a small running service and the numbers you need to work it out.

What You Will Learn

• How to choose SLIs, set SLOs, and use an error budget to balance speed and reliability
• How to design alerts that page on user impact with burn rates instead of noisy thresholds
• A repeatable method for finding the cause of a latency or error spike from metrics, traces, and logs
• How to run an incident: roles, mitigation first, and clear communication
• How to write blameless postmortems whose action items prevent repeats
• Why overload spreads through a system, and how load shedding and degradation stop it
• How canaries, progressive rollout, and fast rollback make changes safe
• How to forecast capacity, keep headroom, and know where autoscaling falls short
• How to set RPO and RTO and prove that backups actually restore
• Optional: measuring toil and deciding what to automate
• Optional: running chaos experiments and game days safely

How This Course Works

Each section starts with the obvious approach, such as alerting when CPU passes 80%, retrying every failure, or deploying to all servers at once. Then it shows the concrete incident where that approach fails, and adds the next idea only when that incident calls for it.

Reliability is about numbers, so targets, budgets, burn rates, and headroom are always worked through with an example. Rate limiting, retries, and failover architecture are covered in the System Design courses, and this course links to them instead of repeating them.

Who This Course Is For

This course is for SREs, infrastructure and platform engineers, and backend engineers preparing for SRE and production-engineering interviews. You should know the basic parts of a web service, such as a load balancer, a database, and a queue, and be able to read a latency graph. You don't need prior on-call experience.

This is not a course on a specific monitoring or orchestration tool, on Linux troubleshooting commands, or on security operations.