DNS failover can direct new DNS lookups away from an endpoint that a health check marks unhealthy. It cannot instantly redirect every visitor: recursive resolvers may still have the old answer cached, and existing connections are not moved to a new server. Understanding that distinction is key to designing a useful failover setup.
What DNS traffic management does
DNS traffic management controls which address an authoritative DNS service returns when a resolver asks for a name. A basic A or AAAA record points to an IPv4 or IPv6 address. A traffic policy can instead select among multiple endpoints using configured weights, geography, or health status.
![Computer Networking Bible: [3 in 1] The Complete Crash Course to Effectively Design, Implement and Manage Networks. Including Sections on Security, Performance and Scalability](https://m.media-amazon.com/images/I/41H4YJnxKgL._SL160_.jpg)
As an Amazon Associate I earn from qualifying purchases.
The resolver asks the domain’s authoritative servers on a client’s behalf and may cache the answer for its time to live (TTL). The client then uses the returned address to make a connection. DNS selects an address; it does not carry the website’s traffic or choose a backend for an already-open connection. The distinction is described in RFC 9199 and RFC 6589.
Windows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallOutdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware matchHow DNS failover works
- A check detects a problem. A monitoring system tests a configured endpoint. What counts as unhealthy depends on the check and provider policy.
- The DNS service changes its answer. If the policy is configured for failover, the authoritative service can stop returning the unhealthy endpoint and return a backup or another healthy target.
- Resolvers refresh their cached answers. A resolver that still has the earlier record may continue returning it until its cache entry expires or it otherwise refreshes.
- A later connection attempt uses the refreshed answer. The client must then make a new connection to the selected target; DNS does not transfer an existing TCP or application session.
Google Cloud DNS documents routing policies and health checks, Amazon Route 53 documents health checks and DNS failover for resources serving the same function, and DNS Made Easy describes monitoring a primary address and switching to a secondary address. These are examples of documented product capabilities, not comparative performance tests: Google Cloud DNS routing policies and health checks, Amazon Route 53 DNS failover, and DNS Made Easy failover configuration.
#1 Best Overall
How long DNS failover takes
There is no single end-to-end failover time. It depends on how quickly the monitoring system detects failure, how the DNS policy reacts, how long resolvers continue using cached answers, and how clients retry or reconnect.
RFC 9199, published by the RFC Editor/IETF in March 2022, gives examples rather than a universal TTL prescription: five minutes may be needed by DNS-based load-balancing or DDoS-prevention users seeking agility; 15 minutes may provide sufficient agility for many operators; and at least one hour is reported as recommended for registry operators and child NS and other records in the cited research context. The RFC cautions that operators must balance agility against other operational benefits. It also notes that some recursive resolvers impose minimum cache times of tens of seconds, so a short TTL does not make all resolvers behave identically. See RFC 9199.
For a provider-specific example, DNS Made Easy’s support article, updated March 11, 2025, describes a two-to-four-minute monitoring window for its configuration. That is a monitoring interval for that service example, not a general DNS failover time or a measured end-to-end recovery guarantee: DNS Made Easy’s configuration article.
Choose a TTL in light of your recovery objective, query load, resolver behavior, record type, and delegation context. The RFC’s authors—G. Moura, W. Hardaker, J. Heidemann, and M. Davids—put the trade-off this way: “There is always a tussle between using shorter TTLs that provide more agility and using longer TTLs that include all the benefits listed above.”
Round robin, health-checked failover, and load balancing compared
| Approach | Health awareness | How it steers traffic | Main limitation |
|---|---|---|---|
| Multiple DNS addresses or round robin | Not inherent; requires separate monitoring and record updates | Publishes several addresses; answer ordering may rotate | Resolvers and clients may handle the addresses differently, and an unhealthy address can remain in use if it is not removed. RFC 1794 and RFC 6589 discuss the relevant DNS load-balancing concepts. |
| Health-checked DNS failover | Checks configured endpoints according to provider policy | Returns a healthy or backup DNS target | Cached answers delay adoption, and DNS controls later lookups rather than existing connections. RFC 9199, Route 53. |
| Geographic or weighted DNS policy | May be combined with checks, depending on implementation | Selects answers by configured weight or location | Resolver location is an estimate of the user’s location, and caching can make steering inexact. Google Cloud DNS, RFC 6589. |
| Load balancer behind a DNS name | Typically selects among backends at the service or network layer | DNS points clients to the balancer, which selects a backend | The balancer is a separate layer and may itself need resilient deployment. RFC 6589. |
Round robin is a distribution technique, not a health check. Publishing multiple addresses alone does not ensure that DNS will stop advertising a failed endpoint. RFC 1794 is an informational historical document; it is useful background on the technique, not a modern product benchmark.
What DNS failover can and cannot protect
DNS failover can help new lookups avoid an endpoint when a configured check identifies it as unhealthy and the policy has a viable alternative. It does not fix the failed application, move an established session, or guarantee that every client reaches the backup immediately. A resolver may still serve a cached address, and client retry behavior also affects recovery.
Rank #4
- Used Book in Good Condition
DNS steering is therefore one part of availability design, not a guarantee that a website stays online through every failure. The backup must be capable of serving the requested function, and the application and its dependencies must also be available there. DNS failover alone does not establish that readiness.
Keep the DNS service reachable too
Protecting web endpoints and protecting authoritative DNS are related but distinct tasks. Traffic steering only helps if resolvers can reach the authoritative service and obtain usable answers. RFC 10001, published in 2026, provides operational guidance for authoritative DNS reachability over IPv4 and IPv6 and calls for DNS-over-TCP availability as a fallback. See RFC 10001.
Changing nameserver delegation to add or switch DNS providers is a separate form of failover. It involves delegation caching and, where DNSSEC is used, coordination of DNSSEC configuration; the mechanisms discussed here do not establish a universal multi-provider recipe.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
