Fundamentals, Terminology, and Core Concepts
Start with a shared vocabulary
Much of this domain is about distinguishing terms that sound interchangeable but are not. A component is a replaceable part; a service is a capability delivered to a consumer; a system is the whole arrangement of components and their interactions. Questions often describe a failure and ask you to classify it, so learn the fault-to-failure chain: a fault is a latent defect, an error is the incorrect internal state it produces, and a failure is the observable deviation from expected behavior. A defect can exist for years without causing a failure.
Quantities candidates confuse
Availability is the fraction of time a service is usable; reliability is the probability it works correctly for a given period; durability is the probability data survives. A service can be available but unreliable (it restarts constantly) or reliable but unavailable during a planned window. Latency is time per request; throughput is requests per unit time; adding concurrency often improves throughput while worsening tail latency. Memorize the arithmetic: 99.9 percent availability over a 30-day month allows roughly 43 minutes of downtime; 99.99 percent allows about 4.3 minutes.
Structural concepts
Know client-server and layered models, stateful versus stateless design, synchronous versus asynchronous communication, and idempotency (repeating an operation yields the same end state). Idempotency is the standard answer whenever a question mentions retries or duplicate messages.
Traps
Questions swap recovery point objective for recovery time, equate monitoring with testing, or ask for the first step when several later steps are also listed. Read for the qualifier: first, best, most likely, primary.
Worked example
A service must be usable 99.9 percent of the month and each request must complete in under two seconds. The first figure is an availability target; the second is a latency target. If a question asks which one is a non-functional requirement, both are; if it asks which drives redundancy design, it is availability.