Regional failover allows you to specify a backup system in the same or a neighboring region that will handle traffic in the event that the primary system fails.
It’s easy and absolutely necessary to have failover configured, no matter what location-based routing service you are using. Since DNS records have no way of knowing if they are pointing to an active or downed system, you need to have an integrated monitoring service configured so users aren’t pointed to downed systems.
How It Works
Failover uses integrated health checks that monitor the status of your systems. If a system fails, your records will automatically update to point to a healthy backup system.
Failover is compatible with ALL of our GeoDNS services.
If you are unfamiliar with GeoDNS, we recommend that you take a look at this article and then come back.
Regions are defined by different GeoDNS services. The simplest service, regional traffic director, uses five predefined regions. Whereas IP Filters use user-defined regions that can consist of continents, countries, cities and/or geographical coordinates.
#1 Regional Traffic Director + Failover
This method uses GTD, our regional traffic routing service. You can learn how to set up GTD here.
When you enable the GTD, it will split your domain’s records into five regional groups. This allows you to create different records for each region.
When a user queries a GTD-enabled domain, Constellix answers with the record that corresponds to their region.
But what happens if the record points to a dead endpoint? Nothing. That’s because you don’t have any monitoring or failover rules configured for that endpoint.
Failover is just as easy to configure with a GTD record as it is for a regular record. You just have to set it up for every GTD record. This could mean up to 5 records, one for each region.
Just enable failover and specify your backup endpoints. You will need to create monitoring checks for each endpoint, including the primary one.
Rinse and repeat for every regional record.
Keep in mind, backup systems can be in other regions. Say for you have five systems, one in each region. You can create failover configurations in each regional record that points traffic to the next closest region in the event of an outage. If the backup system is down, that record will also have a failover rule that points to another region.
The logic would look something like this: Oceania => Asia-Pac => US West => US East => Europe.
#2 IP Filters + Failover
IP filters use an external database to map IP addresses to their originating locations. In Constellix, we use IP filters to funnel traffic to different resources. Conversely, you can also use filters to block traffic from a location.
When you apply an IP filter to a record, you are required to also have a default or World record with the same name. This record is used to answer any queries that don’t match the IP filter. This will be important later.
You can specify backup IP addresses or hostnames for any record with an IP filter applied. Just enable failover, specify your backup systems, then create and apply the associated monitoring checks.
Rinse and repeat for your default/World record.
Are there any other services that you want to use with Failover? Let us know and we’ll write a tutorial for it!