Microsoft Azure team has released a useful tool for Azure Kubernetes Service (AKS) called Azure Monitor: A container health monitoring tool. It’s right now in public preview and the developers are welcome to work on it. This enables to developers to track the health and performance of your Azure Kubernetes Service (AKS) cluster. As it is said, Microsoft will release new features and updates to on-boarding and the portal experience later.
Nodes have a lifecycle of fail and reboot. Therefore it is important that things are kept under watch. Not that they fails almost, but there can be instances. Azure Monitor assures such instances are monitored by collecting the data from processors and memory allocations. Not just them, the logs are also monitored and kept in places. There is a Log Analytics agent for Linux available for collecting purposes. Those logs are stored in the Log Analytics workspace.
What are the information provided?
Azure Monitor for Containers come with in-built set of views that showcase the performance heal of a cluster at any given time or in history.
Identify AKS containers that are running on the node and their average processor and memory utilization, which can help in identifying resource bottlenecks.
Identify where the container resides in a controller or a pod. This knowledge can help you view the controller's or pod's overall performance.
Review the resource utilization of workloads running on the host that are unrelated to the standard processes that support the pod.
Understand the behavior of the cluster under average and heaviest loads. This knowledge can help you identify capacity needs and determine the maximum load that the cluster can sustain. "
Quoted from: Azure Blog
Enhancements to on-boarding
- Using Azure CLI, users can measure the container health with existing Log Analytics workspace
- By clicking UI portal button, will be quickly enable the existing AKS environment.
Usability improvements in the performance grid
- New Cluster tab: cluster performance line chart view for aggregated cluster performance and pod statuses
- Added filter to exclude kube-system from the UI to present customer workloads only
- Added Average/Min/Max/50th/90th/95th percentiles selector for metrics on performance grid and line chart
- Ability to search by name and sort by all columns
Detailed article can be found on Azure Blog.