OpsRamp, a provider of a cloud service that automates IT operations using machine learning algorithms, has added support for Kubernetes clusters running in on-premises or in a public cloud.
Mahesh Ramachandran, vice president of product management for OpsRamp, says that while the percentage of IT environments employing Kubernetes today is relatively small, significant momentum is building around deploying cloud-native applications on the platform.
The OpsRamp Winter Release now can monitor nodes and containers for each cluster, including the ability to provide a breakdown of pods by namespace and resource trends for each Kubernetes cluster, says Ramachandran. OpsRamp can now collect, aggregate, correlate and escalate events from AWS services such as AWS Health, ECS, Redshift, Data Migration Services and CloudWatch.
While many Kubernetes deployments first occur in silos, Ramachandran says IT organizations eventually will incorporate Kubernetes management within the context of more comprehensive management frameworks such as OpsRamp. These organizations have made massive investments in previous generations of platforms that won’t be eliminated anytime soon, he notes.
To simplify the management of IT across all those platforms, the OpsRamp Winter Release also adds topology mapping tools to establish relationships between application components and IT infrastructure for more than 40 enterprise applications. OpsRamp also discovers virtual machines, hypervisor servers and clusters and their relationships to one another within VMware vSphere and kernel-based virtual machine (KVM) environments. OpsRamp has also enhanced the user interface for the service maps it provides to highlight issues associated with a specific IT service outage.
Finally, OpsRamp has updated OpsRamp OpsQ, the event management engine it makes available to further drive AIOps. Automatic incident creation and routing using alert escalation policies is now enabled to to auto-assign incidents based on prior alerts, incident and notification data. Machine learning algorithms now escalate alerts based on learned patterns informed by, for example, business impact, urgency or frequency. OpsRamp OpsQ also now allows IT organizations to supplement the data being used to drive the AI models created by OpsRamp.
Ramachandran concedes it may still be a while before most IT organizations are comfortable enough with AI to manage IT operations. But the end goal is not to replace IT staff, he notes; the goal is to eliminate as much of the noise being generated across the IT environment as possible. Once that noise is eliminated, it then becomes possible for IT administrators to manage IT environments at much higher levels of scale.
The rise of microservices-based applications that rely on ephemeral containers running on Kubernetes cluster is only going to make it more obvious there is a greater need to rely on AI to manage IT at scale, adds Ramachandran. In fact, he notes the amount of data required to automate IT using AI is one more reason IT management frameworks now need to be deployed in the cloud. There simply isn’t enough local server capacity to store all the data needed to drive an AI model, he says.
There is no doubt AI will soon play a much bigger role when it comes to managing IT, and chances are one of the first places those AI models will manifest themselves is in cloud-native computing environments.