How does the DNS Failover system determine if my site is down?

Hosted DNS
NOTE: This FAQ is for the legacy DNSv2 platform and not the new and current DNSv3 platform.

If you are using the HTTP method to monitor your website using the failover system, the monitoring service will look for the HTTP information codes coming back from your web server. For example, if it receives a 200 code back (which means OK), then it will consider the website to be up and ok. It really doesn't matter what content your page sends back.

If your site sends back any of the error codes, such as a 301 (redirect), 403 (forbidden), 404 (not found), 408 (timeout), 500 (internal server error) or any other type of error, it will consider the response a failure and will change the 'A' record to one of your other IP addresses (if you are not using round robin) or simply remove that 'A' record for that particular IP address (if you are using round robin).

It is important to note that the failover service will NOT fail over to one of the other IP addresses if they are not currently reporting a "good" status. If your second IP is also in a failure state, the service will not change your 'A' record because doing so would not improve the situation for your end-users. If, however, one of the secondary IPs did come back while the primary is still down, it will change over.

For other monitoring types such as Generic TCP, SMTP etc. the service simply looks for a response on that port. If the port is closed, or does not respond within the time period you specify, it will be considered bad and will invoke a failover situation.