How do we measure the site availability?

Question

To measure the availability of a web site / API, should the dependencies also be considered? For instance, assume the payment service is down; but the shopping site is still available. Here the customer is not able to complete the purchase since the payment is down. Does this affect the availability measure? Another example is when few of the customers are affected, but the site is still available to other users. Does this affect the availability of the site?

Answer 1

Usually when you break down overall availability metric to individual microservice it is more effective to track and optimize.

Let's take the example of an e-commerce site with a payment microservice, two mostly adopted ways for measuring availability,

Time based availability = (Uptime) / (Uptime + Downtime)

Downtime - Duration of how long the payment microservice was completely unavailable to its consumer (website or another microservice)

Aggregate availability = (Successful Requests) / (Total Requests)

Aggregate availability takes into account all the request that the service failed to service.