Datawire has added a Service Preview to a platform it created to manage application programming interfaces (APIs) for load balancing and ingress to a Kubernetes cluster.
The Ambassador Edge Stack was launched late last year. The Service Preview capability will make it possible for developers to preview changes they want to make to an application before they package up those changes in a container. That Service Preview capability is based on open source code Datawire developed and donated to the Cloud Native Computing Foundation (CNCF). The CNCF makes that code available currently in the form of a sandbox project known as Telepresence.
Service Preview is designed to allow developers to debug their application on their local computer as if it were in their cluster by examining results through the curl command. Service Preview employs fine-grained L7 routing enabled by the Ambassador Edge Stack to send test traffic requests into the development cluster and have those requests routed to and from their local machine. Developers can test treat the local version of the microservice as if it were on a shared cluster, in addition to testing interconnections to adjoining microservices and data stores.
Datawire CEO Richard Li says that while there is a massive amount of interest in Kubernetes, one of the issues organizations face once they adopt it is developer productivity drops, as Kubernetes and containers require developers to adopt a new set of workflow processes. The Service Preview capability makes it possible for developers to preserve their existing workflow processes by enabling them to preview code before it needs to be packaged in containers and then pushed to the Kubernetes cluster.
That approach also minimizes the impact those updates would have on other developers using the same cluster because each developer is sharply reducing the number of times they need to commit code to the Kubernetes cluster, adds Li. The Datawire approach also helps minimize the number of Kubernetes clusters deployed because it serves to increase the utilization efficiency of each cluster, he says.
Of course, IT teams could give each developer their own copy of the development environment. However, Li notes, that approach results in a lot of duplication across each of those environments. In addition, those local development environments are prone to crashing when they run out of memory and over time they don’t tend to effectively reflect the actual production environments on which code will be ultimately deployed.
It’s not clear what level of enthusiasm there is yet for the Telepresence sandbox project being advanced by the CNCF. If ultimately successful, it usually takes a few years for a sandbox project to graduate to the point which it becomes a full-fledged member of the project supported by the CNCF. The challenge is, given the rate at which microservices-based applications are being developed on Kubernetes, the need for a better way to manage the test and development process is becoming more acute with each passing day.