As organizations have started moving more and more of their workloads to Kubernetes, many have added more Kubernetes clusters to their platforms. This introduces the need to manage anywhere from hundreds to thousands of clusters efficiently. However, there is generally no programmatic way to place workloads into the clusters intelligently.
Organizations are looking for a smarter orchestration engine above all these clusters to help facilitate scheduling workloads intelligently on to different clusters to maximize resource usage, based on heuristics such as cost, availability of resources, and so on.
To address these needs, we’re extending Fleet workload placement to schedule workloads to clusters based on new heuristics such as cost and availability of resources.
For added flexibility, we’re providing the option to customize cluster-specific resources by targeting groups of clusters through resource override.
With Azure Kubernetes Fleet Manager, its workload placement experiences, and especially the use of cost and resource availability heuristics for these scheduling decisions, organizations can extract as much utilization as possible from their Kubernetes platforms.
These workload placement/scheduling components are being developed in open source and we are planning to more broadly engage and collaborate with the community via this open-source project.
