Amazon Route 53 health checks are independent resources that can be used by most routing policies when defining a record.  When you create a health check, Route 53 sends requests to the endpoint every 30 seconds, and based on the responses, Route 53 decides if the endpoint is Healthy or UnHealthy and uses that information to determine what value to provide as an answer to the query. 

You can also configure a health check for other “health checks” allowing you to independently verify different tiers of your application before the actual total application is considered healthy. Amazon Route 53 adds up the number of health checks considered healthy and compares that number to the health threshold value you specify. 

With Route 53 health checks you can also monitor the state of a cloud watch alarm. The health check status is healthy when the alarm is in the OK state.  The health check status is unhealthy when the alarm status is in the ALARM state. You can also choose what the health check status is when the alarm is in the INSUFFICIENT state.  The options are healthy, unhealthy or “last known status”. 

When Route 53 receives a query it chooses a record based on the routing policy, it then determines the current health of the selected record by checking the status of the health check for that record and responds to the query with the value of a healthy record.  Unhealthy records are not considered.  If you do not associate a health check with a record, Route 53 treats those records as always healthy. 

The health check is performed by a fleet of health checkers located worldwide. You can use the list of recommended health checkers by region or customize the list to the regions specific to your business.  Health checks are performed every 30 seconds unless you specify every 10 seconds. 

Endpoint health checks can be specified by IP address or by domain name. The health check protocol can be TCP, HTTP, or HTTPS. For the HTTP-related protocols, you can use an optional string matching where you indicate that Route 53 is to search the response body for the string specified. Route 53 considers the endpoint healthy only if the string specified appears entirely within the first 5120 bytes of the response body.   

Finally, for all health checks, you can choose to get notified when it fails.