Management of periodic requests for compute capacity

摘要

A system for monitoring incoming code execution requests and scheduling corresponding code runs is provided. The system may be configured to maintain a plurality of virtual machine instances. The system is further configured to determine whether at least a portion of the incoming code execution requests exhibits periodicity, such that a reduced amount of idle state computing capacity is maintained on the virtual compute system. The system may be configured to add additional compute capacity just before additional code execution requests are expected to be received.