[Review]: Introducing Windows Server System Insights
Windows Server System Insights
As an IT admin, one of the responsibilities you have is to ensure systems continue to run smoothly. That is true for a number of activities and components, such as monitoring if a disk is going to run out of space, determining how much memory and processing a Hyper-V host is consuming so you can plan for new VMs, and many other examples.
System Insights is a new feature available today in the Windows Server 2019 preview that brings local predictive analytics capabilities natively to Windows Server. These predictive capabilities, each backed by a machine-learning model, locally analyze Windows Server system data, such as performance counters and events, providing high-accuracy predictions that help you reduce the operational expenses associated with reactively managing your Windows Server instances.
Because each of these capabilities runs locally, all your data is collected, stored, and analyzed directly on your Windows Server instance, allowing you to use predictive analytics capabilities without any cloud connectivity. In Windows Server 2019, System Insights introduces a set of capabilities focused on capacity forecasting, predicting future usage for compute, networking, and storage.
Current and Upcoming Features
With the current Windows Server 2019 preview build, System Insights users can:
- Browse through predictive capabilities, and either invoke a capability on-demand or configure it to run it on a periodic schedule.
- Visualize prediction outcomes to intuitively understand capacity consumption trends.
- Set custom remediation jobs to automatically run after a capability generates a specific result, helping users automatically mitigate the issues detected by the predictive capabilities.
- View and understand how capacity predictions from an individual Windows Server are trending over a period of time.
- Use PowerShell on remote instances to aggregate prediction outcomes reported by a fleet of related Windows Server instances – e.g. cluster, application tier, rack, and data center – to understand how the fleet overall is trending along compute, storage, or network capacity dimensions.