Load and Failure for Site Reliability Engineers
Load and failure covers how traffic is distributed and what happens when systems are pushed past their limits. It includes load balancing strategies, handling overload through load shedding and graceful degradation, and understanding cascading failures where one component failing overloads its neighbours. It is central to building services that degrade rather than collapse.
Stage: Engineering for Reliability · Part of the Site Reliability Engineer roadmap · Last reviewed 19 November 2026
What to learn in Load and Failure
- Compare round robin, least connections and consistent hashing for balancing load
- Apply load shedding and rate limiting to protect a service under overload
- Use timeouts, retries with backoff and jitter without causing retry storms
- Explain how a cascading failure starts and how circuit breakers stop it
- Degrade gracefully by serving reduced results when a dependency fails
Load and Failure interview questions
What is a retry storm and how do you avoid it?
When many clients retry failed requests at once, they multiply load on an already struggling service. Use exponential backoff with random jitter, cap total retries, and set retry budgets so retries stay a small fraction of traffic.
What is a cascading failure?
A failure that spreads: one component slows or fails, its load shifts to peers that then overload, and they fail too. Defences include load shedding, circuit breakers, sensible timeouts and capacity headroom for the loss of a node.
Where Load and Failure fits in the Site Reliability Engineer roadmap
Load and Failure belongs to the Engineering for Reliability stage (advanced level), which takes about 40 hours. Make systems resilient.
Other units in this stage
Back to the full Site Reliability Engineer roadmap