Amazon Route 53 routing policies: SAA-C03 scenarios

Seven routing policies, one requirement sentence apiece — and the health checks that decide a failover. Match the policy to the need, SAA-C03, 2026 edition.

Try 7 free Route 53 routing-policy scenarios in the exam's own style — every answer explained, no sign-up.

Straight to the 7 free questions ↓

Route 53 questions on the SAA-C03 are rarely about DNS mechanics. They hand you a requirement — a percentage split, a residency law, an automatic regional switch — and ask which of seven routing policies satisfies it. The skill is matching the requirement sentence to the one policy built for it, and knowing why the other six miss. The diagram above shows the same hosted zone answering a single query four different ways depending on the policy in play.

The two “no decision” policies

Simple routing is the default and the baseline: one record, no conditional logic, no health checks. When the stem describes a single resource and asks nothing about choosing between answers, that is your cue. Multivalue answer routing looks similar because it returns several records, but it adds per-record health checks so failing servers drop out of the answer set. It spreads clients across a small fleet without a load balancer — useful, but explicitly not a replacement for one. The exam likes to offer both where only the health-check requirement separates them.

The three “route by who or where” policies

Weighted routing splits traffic by numeric proportion — a 70/30 A/B test, or a tunable shift between two live stacks. The trigger is a controllable percentage. Latency-based routing answers with whichever Region resolves fastest for that user, which may not be the nearest one. Geolocation routing answers by the user’s country or continent and supports a default record for the rest — the right call when the driver is data residency or compliance rather than speed. Latency versus geolocation is the classic trap: read whether the requirement is performance or location.

Geoproximity: the dial

Geoproximity routing also reasons about geography, but its distinguishing feature is the bias — a value that expands or shrinks the geographic area routed to a resource. When a scenario wants to pull more traffic toward an under-used Region without hard-coding percentages, that adjustable catchment is geoproximity, not weighted routing.

Failover: the policy the health check drives

Failover routing is active-passive. You designate a primary and a secondary and attach a Route 53 health check to the primary. While the check passes, Route 53 returns the primary; when it fails, it returns the secondary — automatically, with no manual step. This is the only policy in the set whose answer is decided by a health check, which is why the diagram marks the failover branch with the health-check symbol. If a stem asks for an automatic switch to a standby “with no human in the loop,” the health check is doing the work. Health checks can also be attached to weighted and multivalue records to remove unhealthy targets, but only failover uses one to choose between a primary and a secondary.

Matching method

Work every scenario the same way: find the requirement word and let it name the policy. “Percentage” or “A/B” means weighted; “lowest latency” means latency-based; “by country” or “residency” means geolocation; “expand the area” means geoproximity’s bias; “automatic switch to standby” means failover; “several healthy IPs” means multivalue; and a lone resource with no decision means simple. For how failover fits a wider recovery plan, see disaster recovery strategies and the Design Resilient Architectures domain.

Answer each of the seven below before opening the explanation, and when you miss, pin down the requirement word that decided it. Every explanation names that constraint and takes the plausible-but-wrong policies apart.

An Amazon Route 53 hosted zone receiving a DNS query and dispatching it by routing policy to four different answers. A weighted branch splits a configurable proportion of traffic between two stacks; a latency-based branch selects the Region with the lowest latency; a failover branch, marked with a health-check symbol, sends traffic to a passive secondary only when the primary's health check fails; and a geolocation branch answers by the user's country. The diagram shows that only the failover branch consults a health check to choose its answer.

7 free SAA-C03 practice questions

Answers and explanations — no email wall

SAA-C03 Question 1 of 7

An internal reporting tool runs on a single EC2 instance behind one Elastic IP. There is no second Region, no traffic split and no health-based logic — the team just needs the name reports.example.com to resolve to that one address. Which Route 53 routing policy fits?

Answer: A — Simple routing, mapping the record to the single IP address.

Simple routing is the default: one record, one (or more) values, no conditional logic and no health checks — exactly right when there is a single resource and no requirement to choose between answers. Why not the others: failover routing needs a primary and a secondary record plus a health check, which do not exist here. Multivalue answer routing is for returning several records; with one target it adds machinery for nothing. A single weighted record at weight 100 works mechanically but weighted routing exists to split traffic across records — using it for one target signals the wrong intent. When nothing in the stem asks you to decide between endpoints, the answer is simple routing.

SAA-C03 Question 2 of 7

A company runs two complete, independent versions of a checkout service and wants to send a controllable share — say 70% to the current stack and 30% to a redesigned one — to compare conversion, adjusting the ratio over the trial. Which routing policy does this?

Answer: A — Weighted routing, assigning a numeric weight to each stack's record.

Weighted routing attaches a weight to each record and hands out answers in that proportion, so a 70/30 split you can retune is precisely its job — the standard way to run an A/B comparison or shift traffic between two live environments. Why not the others: latency-based routing chooses by measured network latency, not a ratio you set. Geoproximity routes by geographic distance with a bias, which moves a boundary on the map rather than splitting a defined percentage regardless of location. Multivalue answer returns several records at random without a controllable proportion. A tunable percentage split between two endpoints is weighted routing.

SAA-C03 Question 3 of 7

A stateless API is deployed identically in three Regions. The requirement is that every user is served from whichever Region gives them the lowest network latency at resolution time — which may not be the geographically nearest one. Which routing policy meets this?

Answer: A — Latency-based routing, which answers with the lowest-latency Region for the resolver.

Latency-based routing uses Route 53's latency measurements between networks and Regions and returns the record for the Region that resolves fastest for that user — the direct answer to 'lowest latency, not necessarily nearest.' Why not the others: geolocation routes by the user's country or continent, which can send someone to a nearer-on-the-map but slower Region. Geoproximity also reasons about distance, not measured latency. Weighted routing ignores performance entirely and just splits by proportion. When the requirement word is latency, the policy names itself.

SAA-C03 Question 4 of 7

A payments platform runs an active stack in one Region and a warm standby in a second Region. The requirement: if the active stack stops responding, DNS must automatically send users to the standby with no manual change. How should Route 53 be configured?

Answer: A — Failover routing with a health check on the primary record and the standby as secondary.

Failover routing is active-passive by design: you mark a primary and a secondary and attach a Route 53 health check to the primary. While the check passes, Route 53 answers with the primary; when it fails, Route 53 answers with the secondary — automatic, no human in the loop. Why not the others: a 0% weighted record never receives traffic even when the primary is down, so failover never happens. Latency-based routing splits by performance and would keep sending users to the faster Region even while it is failing. Simple routing returns both addresses blindly with no awareness of health. Automatic active-passive switching is failover routing, and it is the health check that drives it.

SAA-C03 Question 5 of 7

A media service must keep each country's user data on infrastructure inside that country's Region to satisfy data-residency law: German users to eu-central-1, and everyone else to a default endpoint. The rule is about the user's location, not performance. Which policy fits?

Answer: A — Geolocation routing, with per-country records and a default record for the rest.

Geolocation routing answers based on the country or continent the query comes from, and it supports a default record for locations you have not mapped — exactly what a data-residency rule needs: bind Germany to eu-central-1 and catch everyone else with the default. Why not the others: latency-based routing optimises speed and could place German data outside Germany when another Region is faster, breaking the legal requirement. Geoproximity shifts geographic boundaries with a bias but is not built around discrete country-to-resource rules with a catch-all default. Failover is about health, not location. A residency or compliance rule tied to the user's country is geolocation routing.

SAA-C03 Question 6 of 7

Traffic is served from two Regions using geographic distance, but one Region is over-provisioned while the other sits near capacity. The team wants to shift a larger geographic share toward the under-used Region without switching to fixed percentages. Which policy allows this?

Answer: A — Geoproximity routing, using a bias to expand the under-used Region's catchment.

Geoproximity routing sends users to the nearest resource by geographic distance and lets you apply a bias that expands or shrinks the area routing to each resource — raise the bias on the under-used Region and its catchment grows, pulling more traffic in without you defining fixed percentages. Why not the others: weighted routing would hard-code a percentage, which is exactly what the requirement excludes. Geolocation could be hand-tuned country by country but is brittle and is meant for discrete location rules, not a smooth distance-based shift. Latency-based routing optimises measured speed and gives you no lever to rebalance capacity. Expanding or shrinking a geographic catchment with a dial is geoproximity's bias.

SAA-C03 Question 7 of 7

A small fleet of web servers has no load balancer in front of it. The team wants DNS to return several server IPs so clients can spread across them, and to stop handing out any address whose server is failing its health check. Which routing policy is the simplest fit?

Answer: A — Multivalue answer routing, returning up to eight healthy records with health checks.

Multivalue answer routing returns multiple records (up to eight healthy ones) in response to a query and, crucially, checks the health of each so it only returns addresses that are up — a lightweight way to spread clients and route around a dead server without a load balancer. Why not the others: simple routing can list several IPs but performs no health checking, so it will keep returning a failed server's address. Failover routing is active-passive for two endpoints, not a spread across many. Weighted routing distributes by proportion and is meant for controlled splits, not a health-filtered set of equal peers. Note it improves availability but is not a substitute for a real load balancer. Several healthy records returned together is multivalue answer routing.

That is exactly how every question in the course works — answer, explanation, why-not. The real set continues in the practice player: 10 free questions, no sign-up.

Continue with the 10 free questions →

Those 7 questions were the start.

The exam does not test whether you recognise a term — it tests whether you can rule out three plausible answers under time pressure. That is what the explanations above are for, and there are 780 more questions built exactly like them.

Collecting questions yourself

Scattered across forums, of unknown age
Answer keys without reasoning
No idea which domain you are weak in

Practising with a system

780 questions in 12 full tests, SAA-C03 (2026)
Every option explained — including the wrong ones
Readiness per exam domain, and drills for your weakest

Start free — 10 questions

Straight into the player. No account, no email.

Frequently asked questions

Latency-based versus geolocation routing — how does the exam tell them apart?

By the requirement word. Latency-based routing optimises for the lowest measured network latency and can send a user to a Region that is not the nearest on the map. Geolocation routing answers by the user's country or continent — the right choice when the driver is compliance, data residency or serving localised content, where the user's location, not speed, decides.

Does Route 53 failover routing require health checks?

Yes. Failover routing is active-passive and depends on a Route 53 health check attached to the primary record. While the check passes, Route 53 returns the primary; when it fails, Route 53 returns the secondary. Without a health check there is nothing to signal the switch, so the failover would never happen automatically.

Are these real exam questions?

No. They are original scenarios written for this page in the exam's style and difficulty. Reproducing real exam items violates the AWS certification agreement and can cost a candidate their certification — ours come with the reasoning attached instead.

Updated for SAA-C03 (July 2026). The sample questions above are our own work in the style of the exam — not real exam items. The exam itself is set and marked by the certification body.