intermediate

Data fetching state

Model server-state caches through freshness, invalidation, optimistic updates, and background refetching.

Server-state libraries model remote data as a cache with explicit freshness rules — not as a manual Redux slice per endpoint. Concerns include keys, stale time, background refetch, mutations, invalidation, and optimistic UI.

Subtopics: TanStack Query, SWR, cache invalidation, optimistic UI.

On interviews: describe stale-while-revalidate; when to refetch on focus versus poll.

Common pitfalls: using fetch caches for UI toggles, missing invalidation after mutations, and ignoring error/retry semantics.

Checklist:

Cache keys represent query identity.
Separate loading metadata from domain slices.
Invalidate or update after writes.
Plan optimistic paths with rollback.