APIs have become a bedrock part of mainstream enterprise computing architectures. That means they need the same degree of management and monitoring as standard full-blown applications.
A well-functioning API effort needs "advanced monitoring, metrics, and analytics to inform all of the processes-capabilities that not only capture raw traffic data, but also make that data actionable to help prevent abuse, provide insight into developers' experiences, shape product iteration, align internal stakeholders, and shine lights on untapped opportunities," a recent ebook[1] from Google Cloud and Apigee relates.
This is becoming especially important with the increasing numbers of non-developers who are also dabbling in API creation and deployment. A recent survey[2] by Postman finds 53% of those working with APIs do not have the title of "developer" -- up from 41% the year before. With so many people across the enterprise with their hands in APIs, an API program could go quickly off the rails. The business may end up supporting -- and paying for -- dozens of clunky and underused APIs, while taxing its infrastructure with other APIs that are poorly designed and overused.
This calls for greater visibility into the API experience. Enterprises need "insight into which APIs are being adopted and what that adoption may portend about emerging business opportunities or investment priorities," the Apigee ebook's authors urge. "They need to understand how legitimate API traffic differs from traffic from bad actors-and how to thwart the latter without disrupting the former. They need data to align business and IT leaders around the ways developer activity contributes to revenue."
Measurable metrics are always a good thing, and the challenge is identifying and capturing the right metrics for APIs, which tend to operate under new sets of rules that may differ