Profiling workloads in host systems allocated to a cluster to determine adjustments to allocation of host systems to the cluster

1. A computer program product for gathering information on workload processing to allocate computing resources to workloads in a computing system, wherein the computer program product comprises a computer readable storage medium having computer readable program instructions executed by a processor to perform operations, the operations comprising:

updating an aggregate completion duration for workloads that have completed processing;

updating an aggregate consumed resources, comprising an aggregate of resources consumed by the workloads during processing time;

calculating a ratio between the aggregate completion duration and a requested completion duration of a specified workload; and

determining resources to provision for the specified workload by multiplying the ratio and the aggregate consumed resources.

2. The computer program product of claim 1 , wherein the workloads used to update the aggregate completion duration and the aggregate consumed resources comprise both running and completed workloads.

3. The computer program product of claim 1 , wherein the updating the aggregate consumed resources comprises:

determining whether resource consumption, by a running workload of the workloads, is available; and

updating the aggregate consumed resources with the resource consumption of the running workload.

4. The computer program product of claim 1 , wherein the workloads are of a workload class of a plurality of workload classes, wherein each workload class has shared characteristics and processing behaviors, wherein an aggregate completion duration and aggregate consumed resources are separately calculated for the workload classes.

5. The computer program product of claim 4 , wherein the operations further comprise:

determining whether there is an aggregate completion duration and aggregate consumed resources for a workload class when processing a first workload of the workloads of the workload class; and

generating the aggregate completion duration and the aggregate consumed resources based on the first workload in the workload class in response to determining that there is no aggregate completion duration and aggregate consumed resources in the workload class based on a completion duration and consumed resources for the first workload.

6. The computer program product of claim 1 , wherein the operations further comprise:

determining whether an aggregate consumed resources is provided for one of the workloads; and

using a default value for the aggregate consumed resources in response to determining that the aggregate consumed resources is not provided, wherein the default value is based on a number of resources to process a workload within the requested completion duration.

7. The computer program product of claim 1 , wherein the aggregate completion duration for a workload that has completed processing comprises:

a difference between a completion time and submission time for the workload in response to the completion time and the submission time being available for the workload; and

a difference between a current time and the submission time in response to the completion time not being available for the workload and the workload is for a first time noticed to be in a completed state since the determining the workloads.

8. The computer program product of claim 1 wherein each of the resources comprises at least one of a processor core, an amount of memory space, an Input/Output capacity, and a network capacity.

9. A system for gathering information on workload processing to allocate computing resources to workloads, comprising:

a processor; and

a computer readable storage medium having computer readable program instructions executed by the processor to perform operations, the operations comprising:

updating an aggregate completion duration for workloads that have completed processing;

updating an aggregate consumed resources, comprising an aggregate of resources consumed by the workloads during processing time;

calculating a ratio between the aggregate completion duration and a requested completion duration of a specified workload; and

determining resources to provision for the specified workload by multiplying the ratio and the aggregate consumed resources.

10. The system of claim 9 , wherein the workloads are of a workload class of a plurality of workload classes, wherein each workload class has shared characteristics and processing behaviors, wherein an aggregate completion duration and aggregate consumed resources are separately calculated for the workload classes.

11. The system of claim 10 , wherein the operations further comprise:

determining whether there is an aggregate completion duration and aggregate consumed resources for a workload class when processing a first workload of the workloads of the workload class; and

generating the aggregate completion duration and the aggregate consumed resources based on the first workload in the workload class in response to determining that there is no aggregate completion duration and aggregate consumed resources in the workload class based on a completion duration and consumed resources for the first workload.

12. The system of claim 9 , wherein the operations further comprise:

determining whether an aggregate consumed resources is provided for one of the workloads; and

using a default value for the aggregate consumed resources in response to determining that the aggregate consumed resources is not provided, wherein the default value is based on a number of resources to process a workload within the requested completion duration.

13. The system of claim 9 , wherein the aggregate completion duration for a workload that has completed processing comprises:

a difference between a completion time and submission time for the workload in response to the completion time and the submission time being available for the workload; and

a difference between a current time and the submission time in response to the completion time not being available for the workload and the workload is for a first time noticed to be in a completed state since the determining the workloads.

14. A method for gathering information on workload processing to allocate computing resources to workloads in a computing system, comprising:

updating an aggregate completion duration for workloads that have completed processing;

updating an aggregate consumed resources, comprising an aggregate of resources consumed by the workloads during processing time;

calculating a ratio between the aggregate completion duration and a requested completion duration of a specified workload; and

determining resources to provision for the specified workload by multiplying the ratio and the aggregate consumed resources.

15. The method of claim 14 , wherein the workloads are of a workload class of a plurality of workload classes, wherein each workload class has shared characteristics and processing behaviors, wherein an aggregate completion duration and aggregate consumed resources are separately calculated for the workload classes.

16. The method of claim 15 , further comprising:

determining whether there is an aggregate completion duration and aggregate consumed resources for a workload class when processing a first workload of the workloads of the workload class; and

generating the aggregate completion duration and the aggregate consumed resources based on the first workload in the workload class in response to determining that there is no aggregate completion duration and aggregate consumed resources in the workload class based on a completion duration and consumed resources for the first workload.

17. The method of claim 14 , further comprising:

determining whether an aggregate consumed resources is provided for one of the workloads; and

using a default value for the aggregate consumed resources in response to determining that the aggregate consumed resources is not provided, wherein the default value is based on a number of resources to process a workload within the requested completion duration.

18. The method of claim 14 , wherein the aggregate completion duration for a workload that has completed processing comprises:

a difference between a completion time and submission time for the workload in response to the completion time and the submission time being available for the workload; and

a difference between a current time and the submission time in response to the completion time not being available for the workload and the workload is for a first time noticed to be in a completed state since the determining the workloads.