Building a container application is one thing; learning to live with it in a production environment on the other hand represents a major set of IT operational challenges. To help address this issue the Cloud Native Computing Foundation (CNCF) this week announced it has formed Project Prometheus to oversee the development of an open source monitoring software optimized for containers.
Project Prometheus is the second major project the CNCF has taken on since the Kubernetes container orchestration project that was originally spearheaded by Google. Prometheus itself is already integrated with Kubernetes.
Alexis Richardson, chair of the technical oversight committee for Project Prometheus at the CNCF, says that with each successive wave of technologies it becomes apparent that IT operations teams need tools that are specifically optimized for each versus trying to extend legacy tools that don’t provide as much insight or actionable intelligence.
The core code base of Project Prometheus was originally developed by SoundCloud, a music web site. That continued development of that code is now being overseen by the CNCF. Richardson is also the CEO of Weaveworks, a provider of container networking software that incorporates Project Prometheus code.
Originally inspired by Google’s internal monitoring tools known as Borgmon, Prometheus has also been employed by Digital Ocean, Ericsson, CoreOS, Red Hat and Google. The end goal is to make sure that various implementation of the Prometheus IT monitoring software actually wind up being interoperable with one another.
Of course, the tradeoff between having monitoring tools optimized for different classes of technologies is that IT operations teams have to master them all. The end result often makes it a challenge the pinpoint the source of particular problem. In addition, most IT operations teams can’t always afford to have a dedicated specialist for each IT monitoring tool, which often results in them having to swivel between various management consoles to figure out what is actually taking place in their environment at any given time.
Nevertheless, Richardson notes the ability to delve into data being generated at high throughput levels using IT monitoring tools such as Prometheus that are designed to support time series queries makes the investment worthwhile.
In general, Richardson says most IT organizations are not yet prepared to support the dynamic nature of microservices architectures that by their very nature are much more fluid than legacy IT environments. As such, Richardson says that at least for the time being there’s a philosophy in place that place the onus for managing the code they write on the shoulders of the developers who actually wrote the code. Whether that changes in time remains to be seen. But for the moment Richardson says the ultimate impact microservices architectures will have on traditional DevOps structures inside more traditional IT organizations is as yet unknown.
In the meantime as part of an effort to increase overall trust in terms of assembling, deploying and managing cloud native applications the folks at CNCF are making it clear that Prometheus and Kubernetes are simple the first the many additional projects to come.