In the good world of Route 53 you have 3 different types of health checks. These are endpoint health checks, calculated health checks, and cloud watch alarm health checks. Now what is a health check for? Well if your endpoint doesn’t return a healthy response then you can have your DNS set to failover to another resource. First off there’s the endpoint health check, this uses TCP, HTTP, or HTTPS. It sends a health check every 30, can be set to 10 for a higher price, seconds from 15 global health check endpoints. If over 18% of the results are healthy then the server is deemed healthy. Next up we have calculated health checks, this is a health check for health checks. You can have one parent health check monitor up to 256 child health checks and use AND, OR, and NOT operators between them to have one health check fail if a certain combination of child health checks pass or not. Finally we have the cloud watch alarm health check, this is especially useful if you want to set your own customer alarm that the health check then monitors rather than the endpoint itself. It can also be useful if you want to run a health check on an endpoint in a VPC as that resource will not have access to the global health check endpoints as they are on the public internet. The cloud watch alarm monitors the health of the endpoint using customer alarms and the global endpoints monitor that cloud watch alarm. I should note as well you will need to allow the IPs of the global endpoints through to your instance in order to run the health check, unless of course you are utilizing cloud watch alarm health check in regards to a private VPC as mentioned before.