In my Service Matters blog post, I discussed the Real Availability and what customers should expect from their cloud providers.
Real Availability is the true measure of customer availability by looking at every incident that results in a customer outage (a Priority 1 or P1 incident). We see four groups of P1s:
1) incidents caused by our infrastructure
2) incidents caused by the software on our Service Automation Platform
3) incidents caused by a third-party provider
4) incidents caused by the customer.
In each of the above categories, we are measuring incidents and not synthetic transactions. We are not looking at error logs or monitoring a ping from an external service to the edge of our network. We use customer P1 incident start and stop times to measure Real Availability. This is the only accurate way to measure the availability experienced by our customers.
What Does Real Availability Look Like?
We publish Real Availability for each customer covering the last 90 days on a Dashboard built within the ServiceNow platform (see graphic below):
P1 customer incidents initiate from multiple sources. One source is when a customer opens a P1 incident on our Customer Service System. Another is when our internal or external third party monitoring detects an event that generates a P1 incident. Note that not all monitoring events result in P1 incidents. But if they do, then the incident start time is set to be the time that monitoring first discovered the event.
The Real Availability calculation that we share with our customers is the actual measurement of how they experience availability. It's not machine-generated by a monitoring system but the real incident time. And, yes, some customers have long-running incidents that are self-inflicted, such as when they have internal network or authentication server issues. We add those incidents to the Real Availability calculation because this is the actual availability of ServiceNow from the customer perspective.
We have seen P1 incidents from customers who have a wireless network access point misbehaving at their company. When the access point has an issue, our customers call us saying that they cannot reach ServiceNow. They don't realize that their own internal network caused the incident.
When the customer calls us, we work with them to determine the root cause of the incident. In this example, even though the root cause is on the customer's wireless network, we still count the outage time associated with this incident when calculating Real Availability. It is the actual availability that the customer is experiencing, regardless of root cause (note that because Real Availability captures P1 incidents it is not used to measure contractual availability or service credits).
Real Availability changes the game for the industry when talking about the customer experience.
