How long a piece of data stays valid before being refreshed or expired, the central cache and DNS lever.
TTL (Time To Live) is the duration after which a piece of data is considered expired and must be refreshed, evicted, or revalidated. DNS records have TTLs (how long a resolver caches them), CDN responses have TTLs (how long they're served from edge), session tokens have TTLs (how long they're valid before re-auth), cache entries have TTLs. Picking the right TTL is a freshness/cost tradeoff: shorter TTLs are fresher but cause more origin load and slower failover.
DNS TTL specifically is one of the highest-impact reliability levers most teams ignore. A 24-hour DNS TTL means even after you flip the DNS to your DR region, customers won't see the change for up to 24 hours. Lowering critical TTLs to minutes (well in advance, not during the incident) is what makes DNS-based failover useful as an emergency tool.
See the part of the platform that handles ttl (time to live) in production.