Call Failover Routing

Overview

Perhaps the most important feature our network offers is a variety of ways to deliver callers to their intended destination. The most effective method to do this is by specifying failover servers in your endpoint groups. You can read all about that here: Endpoint Groups. Alternatively, you can forward calls to an external phone number - which we describe here: Forwarding Phone Numbers

However, this article is about what happens after Skyetel exhausted all available routes and has to fail the call. (sad)

Endpoint Outages

Before making a selection, it is important to make a distinction between a Partial Outage or a Total Outage

Partial Outages

A partial outage is caused by your endpoint not responding to one or two of our regions, but having at least one living region active. This is what a partial outage looks like from our Endpoint Health section.

Screen_Shot_2019-02-13_at_3.54.20_PM.png

Total Outage

A total outage is when ALL of our regions cannot reach your endpoint:

Screen_Shot_2019-02-13_at_3.54.26_PM.png

Picking your Poison

While no one likes to decide how to best end a failed call, we at least have some options.

Screen_Shot_2019-02-13_at_3.48.35_PM.png

Play an error message (Default)

This is our default option. When callers dial a number that is failing to route to an endpoint, they will hear a friendly British lady say that the call is experiencing temporary technical issues. This is the default option because we feel it is the most gentle, and it won't artificially increase your billing. This is the preferred method for users who frequently experience total outages. 

Retry the call from different regions

This is an option for users who are frequently experiencing partial outages of their endpoint, not total outages. However - Failed calls incur additional costs. Failed calls are charged for the Caller ID lookup and Spam Lookup for every time we attempt to deliver the call, even though the call was never delivered to an endpoint. Additionally, because we are relaying the call through a relay server, successful calls are double billed for their usage.

Extra Billing (Please Read!)

One of the foundational philosophies of our network architecture is that we do not link our regions together for any reason at anytime. Each region is 100% siloed from each other. This insures that our network cannot have a cascading outage or a failure in one region cannot cause any disruption in any other region. 

In order to make this feature work, we created a relay server that attempts the call on your behalf through each of our regions. In this way, a call can be retried through our regions without requiring our regions to know about each other's status. However, due to limitations in our routers, our network interprets this behavior as 2 calls (one from the PSTN to the Relay server and then another from the Relay Server to the Endpoint) instead of 1 Call (One from the PSTN to the Endpoint). Because of this, your call may be double billed.

Was this article helpful?
0 out of 1 found this helpful
Return to top