Cloud computing capacity management system using automated fine-grained admission control

A method for automated fine-grained admission control for a cloud computing system, comprising:

• obtaining a predicted deployed resources signal that comprises a plurality of different resource usage predictions corresponding to a plurality of different deployment grains, each deployment grain of the plurality of different deployment grains including a combination of a subscriber identifier, a resource identifier, and a geographic identifier associated with a set of one or more users of the cloud computing system, wherein a resource usage prediction for a deployment grain from the plurality of different deployment grains predicts an amount of cloud computing resources that will be consumed by the deployment grain in an upcoming time period;
• obtaining a capacity constrained regions signal that identifies regions of the cloud computing system where demand is expected to exceed capacity at a future point in time;
• creating a capacity mitigation policy based on the predicted deployed resources signal, the capacity constrained regions signal, a predicted capacity effect associated with an estimated effect on availability of computing resources for the deployment grain as a result of the capacity mitigation policy being applied to the deployment grain, and a predicted user effect associated with an estimated effect on at least some of the users of the cloud computing system as a result of the capacity mitigation policy being applied to the deployment grain, wherein the capacity mitigation policy protects available capacity of the cloud computing system for one or more users associated with the deployment grain;
• and
• enforcing resource limits against incoming resource requests from the one or more users associated with the deployment grain based on the capacity mitigation policy.

A system, comprising:

• at least one processor;
• memory in electronic communication with the at least one processor;
• and
• instructions stored in the memory, the instructions being executable by the at least one processor to:
• obtain a predicted deployed resources signal that comprises a plurality of different resource usage predictions corresponding to a plurality of different deployment grains, each deployment grain of the plurality of different deployment grains including a combination of a subscriber identifier and a resource identifier, and a geographic identifier associated with a set of one or more users of the cloud computing system, wherein a resource usage prediction for a deployment grain from the plurality of different deployment grains predicts an amount of cloud computing resources that will be consumed by the deployment grain in an upcoming time period;
• obtain a capacity constrained regions signal that identifies regions of the cloud computing system where demand is expected to exceed capacity at a future point in time;
• create a capacity mitigation policy based on the predicted deployed resources signal, the capacity constrained regions signal, a predicted capacity effect associated with an estimated effect on available capacity of computing resources for the deployment grain as a result of the capacity mitigation policy being applied to the deployment grain, and a predicted user effect associated with an estimated effect on at least some of the users of the cloud computing system as a result of the capacity mitigation policy being applied to the deployment grain, wherein the capacity mitigation policy protects available capacity of the cloud computing system for one or more users associated with the deployment grain;
• and
• resource limits against incoming resource requests from the one or more users associated with the deployment grain based on the capacity mitigation policy.