What is Kubernetes ?
Modern software development is based on new approaches and toolsets to allow agility and quicker, some say continuous, development and deployment of software applications. Kubernetes (a.k.a K8S) is an open source infrastructure that allows building such applications based on containers, which are micro-services packaged with their dependencies and configurations. Containers (sometimes called pods), are small OS instances that replace the need to setup and keep running full-blown VMs (virtual machines) 24x7. It is still a common practice in order to make sure we have enough computing power to accommodate multiple users that are trying to access our web application during peak hours. In contrast, containers are rapidly started when computing power is required and then automatically terminated when the demand is lower thus generating dramatic savings in the business operational costs.
Get full visibility into Microsoft Azure AKS
Microsoft Azure provides AKS (Azure Kubernetes Services) resources for developers to quickly create their favorite infrastructure. Once their AKS is up and running in production, it is important to make sure it functions properly. This is where the AutoMonX Azure sensor pack comes into play. Our sensor pack already monitors 32 different aspects of Microsoft Azure and thanks to its tight integration with a leading tool like PRTG, most complex Azure infrastructure components can be auto-discovered and monitored in minutes.
A quick view of the AKS cluster can be achieved by looking at five sensors that are automatically created during the auto-discovery and monitoring automation phases.
Service Health - the health of the AKS cluster as reported by Azure
Cluster Metrics - AKS consumption of computing resources
Deployments - How many pods (containers) are currently being active per each your application (deployment)
Internal Deployments - The overhead of the AKS cluster. This is the stuff that needs to run so your AKS will be able to function
Kubernetes Nodes - how many working nodes (workers) are running. Pods need a physical server to run on (just like VMs and ESX). The more pods you need, the more worker nodes would be required to support them.
AKS Cluster Utilization Metrics
For instance, in minutes you can get your AKS cluster utilization metrics such as CPU, Memory, Disk usage, Network statistics and more
And the best part is that now you have history being collected for all those metrics, that can be used for notifications in real-time when things go wrong and for troubleshooting of performance issues
Comments