DNS resolution

DNS maps hostnames to records (A, AAAA, CNAME, MX, TXT, SRV). Resolution is recursive from the client through resolvers to authoritative nameservers. TTL controls cache lifetime; low TTL speeds propagation but increases lookup load and latency variance.

					App → OS stub resolver → ISP/recursive resolver → root → TLD → authoritative NS → answer
				

Node services resolve DNS on every new connection unless you cache at the agent layer (`lookup` options, custom agents, or service mesh sidecars). CNAME chains add round trips. NXDOMAIN and SERVFAIL surface as connection errors that look like application bugs.

On interviews: diagnose "works in browser, fails in Node" as DNS or TLS trust differences; explain TTL trade-offs during migrations; describe negative caching and stale records.

Common pitfalls: hard-coding IPs that change; ignoring IPv6 AAAA when only A is tested; assuming instant global propagation after a record change.

The trade-off is cache freshness versus lookup cost and migration safety.

Checklist:

• Trace recursive resolution steps.
• Explain TTL, CNAME, and authoritative vs recursive.
• Name symptoms of DNS failure in Node logs.
• Plan cutovers with lowered TTL ahead of time.