Migrating workloads between computing platforms according to resource utilization

A system comprising:

• one or more processors;
• and
• one or more computer-readable media storing computer-executable instructions that, when executed by the one or more processors, cause the one or more processors to perform operations comprising:
• analyzing utilization data representing computing resources used by a workload running on a first virtual machine type over a period of time in a service provider network, the first virtual machine type being allocated a first amount of computing resources;
• identifying, based on analyzing the utilization data, a use pattern for the workload that occurred during a portion of the period of time;
• identifying a second virtual machine type that is more optimal for running workloads exhibiting the use pattern than the first virtual machine type, the second virtual machine type being allocated a second amount of computing resources that is different than the first amount of computing resources;
• predicting a future time at which the workload will exhibit usage within a threshold similarity to the use pattern;
• and
• generating a migration schedule to migrate the workload from the first virtual machine type to the second virtual machine type at or before the future time.

A method comprising:

• receiving utilization data representing a resource-utilization pattern according to which a workload utilized computing resources of a service provider network over a first period of time;
• generating, based at least in part on a resource-utilization pattern, a migration schedule according to which a workload is scheduled to be migrated between virtual machine types for a second period of time;
• during the second period of time, and according to the migration schedule:
• migrating the workload to a first virtual machine of a first virtual machine type prior to or during a first timeframe, the first virtual machine being allocated a first amount of first computing resources;
• and
• during or subsequent to the first timeframe, migrating the workload to a second virtual machine of a second virtual machine type prior to or during a second timeframe, the second virtual machine being allocated a second amount of second computing resources that is different than the first amount,
• wherein the second amount of the second computing resources is more optimal for hosting the workload during the second timeframe than the first amount of the first computing resources.