Netflix has been busy lately. They’ve steadily worked to become well-known not just for introducing binge-watching to our lives, but also their engineering team’s efforts to create new open-source software. Their latest tool is Fenzo, an open-source scheduling framework for Apache Mesos.
According to Netflix: “Fenzo enables frameworks to better manage ephemerality aspects that are unique to the cloud. Our use cases include a reactive stream processing system for real time operational insights and managing deployments of container based applications.”
The tool uses a scheduling model focused on large-scale deployments with a variety of tasks and resources, each with a variety of constraints and optimization needs. To save time and to make prioritization easier, tasks in Fenzo have an urgency rating and a fitness factor that determines how well the task fits on a host. Users can also decide at what threshold a tasks fits specific criteria for urgency and fitness, as well as whether the task has hard constraints as an additional filter before it is scheduled. Fenzo also supports cluster autoscaling.