Abstract: Disclosed embodiments provide techniques for load balancing of computer jobs in a distributed computer network. A health score is determined for each node of the pool of nodes, which can include native applications, virtual machines, and/or containers. A future resource availability score is determined for each node of the pool of nodes corresponding with a predetermined future time. A schedule eligibility score is computed for each node based on the health score and future resource availability score. A new job is assigned to the node with the optimal schedule eligibility score. In this way, rather than only considering the current workload conditions of the node, a future resource availability score is computed for each node, and this score is used as a factor in the assigning of jobs to the node. This provides an opportunity for improved resource utilization and improved overall system reliability.

Abstract: Disclosed embodiments provide techniques for load balancing of computer jobs in a distributed computer network. A health score is determined for each node of the pool of nodes, which can include native applications, virtual machines, and/or containers. A future resource availability score is determined for each node of the pool of nodes corresponding with a predetermined future time. A schedule eligibility score is computed for each node based on the health score and future resource availability score. A new job is assigned to the node with the optimal schedule eligibility score. In this way, rather than only considering the current workload conditions of the node, a future resource availability score is computed for each node, and this score is used as a factor in the assigning of jobs to the node. This provides an opportunity for improved resource utilization and improved overall system reliability.