Why systems fail catastrophically under overload instead of degrading gracefully, and the patterns - load shedding, backpressure, circuit breakers - that prevent cascading failure.
How to add features, fix bugs, and make progress in codebases you did not write and do not fully understand - without making things worse.
Why production incidents take longer than the underlying technical problem warrants, what coordination failure looks like, and how structured response shortens the time to resolution.
The three categories of time-related bugs that appear in production - timezone confusion, clock skew across distributed systems, and DST - and how to avoid each.
Why renaming a field breaks integrations you did not know existed, what counts as a breaking change, and how to evolve APIs without causing incidents.
Why feature flags accumulate as technical debt, the failure modes teams hit at scale, and how to manage them so they stay an asset instead of becoming a liability.
Why offset-based pagination produces duplicates and skips at scale, and how cursor pagination fixes both the correctness and performance problems.
What eventual consistency means in practice - the real bugs it causes, why distributed systems accept it, and how to design around it instead of fighting it.
Why schema migrations that work in development cause production incidents, and the expand-contract pattern that makes them safe to run on live systems.
Why retries without idempotency cause duplicate charges, duplicate emails, and duplicate records - and how to design write operations that are safe to retry.
Code coverage measures how much of your code is executed by tests. It is a useful indicator and a dangerous target. The difference between the two determines whether coverage improves your software or just your numbers.
Production traffic will test your system whether you do it first or not. Load testing before launch finds the capacity limits, bottlenecks, and failure modes that only appear under real conditions.
