The Cloud Native Computing Foundation (CNCF) announced today at a PromCon 2018 conference that the open source Prometheus container monitoring software project has officially become the second project to graduate, following the Kubernetes project it also oversees.
Richard Hartmann, a Prometheus team leader system architect for SpaceNet, a provider of online satellite images, says the next major focus for the container monitoring software project will be a focus on developing dashboards for various classes of use cases involving monitoring containers. Those dashboards should help accelerate adoption of container monitoring software that today requires a fair amount of developer expertise to employ, he says. At the same time, Hartmann notes the Prometheus team is also working on developing user stories that will provide examples of how to employ Prometheus.
Longer term, Hartmann says CNCF will think through what will be required to add more resiliency to Prometheus deployments by adding capabilities that make it easier to dynamically scale out as container environments grow.
The CNCF also expects that providers of container monitoring software and services will use Prometheus as the foundation for delivering commercially supported offerings in much the same way open source projects are incorporated into other products, he says.
Prometheus was created by SoundCloud in 2012 and contributed to CNCF in May 2016. The project has had 30 official releases since then and now has nearly 20 active maintainers that have overseen the implementation of 19,000 commits from a base of more than 4,300 contributors.
Hartmann says the rise of containers will almost certainly require organizations to employ monitoring more broadly. Applications based on a microservices architecture built using containers tend to be complex, with the benefit that the applications can be updated easily. But tracking the coming and going of ephemeral containers requires a sophisticated monitoring tool; otherwise, IT operations teams will be blind to any events leading up to failure, he notes.
While most enterprise IT organizations have tools to monitor various aspects of their IT environments, very few of them are making broad use of application performance monitoring (APM) tools mainly because the cost of employing those tools across every application they run would be considerable. But as organizations transition to containerized applications, it’s already clear they will need to apply monitoring more broadly. The CNCF is effectively trying to drive down those costs by making Prometheus monitoring software available as an open source project.
Less clear is the degree to which IT organizations will deploy monitoring software specifically optimized for containers running on Kubernetes versus employing tools capable of monitoring both legacy and containerized applications. While Kubernetes enjoys broad industry support, most IT organizations have only just begun to deploy it in a production environment. Another issue will be deciding whether to deploy those monitoring tools on-premises or rely on a cloud service.
Regardless of the approach chosen, monitoring tools are critical to drive any set of integrated DevOps processes. As more organizations embrace those DevOps processes, the need for sophisticated monitoring tools only becomes that much more apparent.