Route53 Flashcards
What is the difference between a CNAME record and an ALIAS record?
An alias is automatically resolved (and cached) by the DNS.
A CNAME has to be re-resolved by the client (meaning a second request you have to make).
This means that an A record and an ALIAS can be used interchangeably.
How do you get domains for Route53?
You can buy it through a 3rd party registrar and transfer it.
Route 53 is also a domain registrar, so you can buy it through R53 without needing to transfer it.
List the Route 53 routing policies.
Simple, weighted, latency, failover, geolocation, geoproximity, multi-value.
In R53, what is the diff between a geolocation routing policy and a geoproximity one?
Geolocation is based on user location. We specify what IP a country (or continent) should go to, with some default for it.
Geoproximity routes traffic based on the geographical location between users and resources.
How can you shift traffic from one region to another?
Use a geoproximity routing policy and change the bias.
How does a simple routing policy work?
When you want to redirect to a single resource.
You can’t attach health checks to a simple routing policy.
If multiple values are returned, a random one is chosen BY THE CLIENT.
How does a weighted routing policy work?
You control the % of traffic that goes to a specific endpoint. Can be useful for splitting traffic between two regions, or testing a small amount of traffic on a new app version.
How does a latency routing policy work?
One of the most common routing policies. Redirects to server that has the least latency.
Be wary that this can sometimes give you weird things like a user in Germany being directed to the US if that’s the lowest latency for the user.
How does a failover routing policy work?
Primary record is associated with a health check. You can only have one primary, and one secondary.
How do health checks work in Route 53?
After x checks fail it’s “unhealthy” (default 3).
Same for an unhealthy target becoming healthy.
Default health check interval is 30s (“fast health check” at 10s for higher price), but you have 15 checkers so there’s a request every 2 seconds on average.
There’s HTTPS support but no SSL verification.
Can integrate the health check into CloudWatch.
