Peer-based optimal performance configuration recommendation

1. A method comprising:

monitoring, by one or more processors, instances of an application, wherein the instances of the application are running on a plurality of devices;

receiving, by one or more processors, data related to: configuration parameters of the application, system parameters of the plurality of devices, and performance metrics of the plurality of devices;

determining, by one or more processors, correlation and dependencies between the configuration parameters with associated system parameters, using an all pair analysis and a Pearson product-moment correlation coefficient (PPMCC), wherein using the all pair analysis comprises:

testing, by one or more processors, possible discrete combinations of the configuration parameters and the system parameters, using test vectors to parallelize the tests of parameter pairs;

determining, by one or more processors, dependencies between the configuration parameters with performance metrics; and

recommending, by one or more processors, to a user, configuration settings of the application, based on the analysis of the data.

2. The method of claim 1 , further comprising:

registering, by one or more processors, for an account for access by the user; and

receiving, by one or more processors, access to the configuration settings recommended to a plurality of users of the application running on the plurality of devices.

3. The method of claim 1 , further comprising:

subsequent to the recommendation, continuing, by one or more processors, to use a current configuration setting; and

testing, by one or more processors, the recommended configuration settings in a separate device with workload replication.

4. The method of claim 1 , further comprising:

automatically changing, by one or more processors, current configuration settings of the user to the recommended configuration settings.

5. The method of claim 1 , wherein determining correlation and dependencies between the configuration parameters with associated system parameters using the PPMCC comprises:

constructing, by one or more processors, a correlation matrix that includes a covariance of two variables divided by a product of a standard deviation of each of the two variables, wherein the two variables are a value for one of the configuration parameters and a value for one of the system parameters.

6. The method of claim 1 , wherein determining dependencies between the configuration parameters with performance metrics comprises:

building, by one or more processors, a performance model for a given value of one of the system parameters, using a plurality of values for the configuration parameters associated with the given value of one of the system parameters and values for one of the performance metrics by creating a multi-dimensional graph, wherein the configuration parameters, the system parameters, and the performance metrics, each, represent a separate axis on the graph; and

determining, by one or more processors, which value of the plurality of values for the configuration parameters corresponds to one of the values for one of the performance metrics when the value for one of the performance metrics is at a highest point for the given value of one of the system parameters.

7. The method of claim 1 , wherein, subsequent to the recommendation, the user has an option to select from the group consisting of: accepting the recommended configuration settings, rejecting the recommended configuration settings, and modifying the recommended configuration settings.

8. A computer program product comprising:

one or more computer readable storage media and program instructions stored on the one or more computer readable storage media, the program instructions comprising:

program instructions to monitor instances of an application, wherein the instances of the application are running on a plurality of devices;

program instructions to receive data related to: configuration parameters of the application, system parameters of the plurality of devices, and performance metrics of the plurality of devices;

program instructions to determine correlation and dependencies between the configuration parameters with associated system parameters, using an all pair analysis and a Pearson product-moment correlation coefficient (PPMCC), wherein using the all pair analysis comprises:

program instructions to test possible discrete combinations of the configuration parameters and the system parameters, using test vectors to parallelize the tests of parameter pairs;

program instructions to determine dependencies between the configuration parameters with performance metrics; and

program instructions to recommend to a user, configuration settings of the application, based on the analysis of the data.

9. The computer program product of claim 8 , further comprising:

program instructions to register for an account for access by the user; and

program instructions to receive access to the configuration settings recommended to a plurality of users of the application running on the plurality of devices.

10. The computer program product of claim 8 , further comprising:

subsequent to the recommendation, program instructions to continue to use a current configuration setting; and

program instructions to test the recommended configuration settings in a separate device with workload replication.

11. The computer program product of claim 8 , further comprising:

program instructions to automatically change current configuration settings of the user to the recommended configuration settings.

12. The computer program product of claim 8 , wherein program instructions to determine correlation and dependencies between the configuration parameters with associated system parameters using the PPMCC comprise:

program instructions to construct a correlation matrix that includes a covariance of two variables divided by a product of a standard deviation of each of the two variables, wherein the two variables are a value for one of the configuration parameters and a value for one of the system parameters.

13. The computer program product of claim 8 , wherein program instructions to determine dependencies between the configuration parameters with performance metrics comprise:

program instructions to build a performance model for a given value of one of the system parameters, using a plurality of values for the configuration parameters associated with the given value of one of the system parameters and values for one of the performance metrics by creating a multi-dimensional graph, wherein the configuration parameters, the system parameters, and the performance metrics, each, represent a separate axis on the graph; and

program instructions to determine which value of the plurality of values for the configuration parameters corresponds to one of the values for one of the performance metrics when the value for one of the performance metrics is at a highest point for the given value of one of the system parameters.

14. The computer program product of claim 8 , wherein, subsequent to the recommendation, the user has an option to select from the group consisting of: accepting the recommended configuration settings, rejecting the recommended configuration settings, and modifying the recommended configuration settings.

15. A computer system comprising:

one or more computer processors, one or more computer readable storage media, and program instructions stored on the computer readable storage media for execution by at least one of the one or more processors, the program instructions comprising:

program instructions to monitor instances of an application, wherein the instances of the application are running on a plurality of devices;

program instructions to receive data related to: configuration parameters of the application, system parameters of the plurality of devices, and performance metrics of the plurality of devices;

program instructions to determine correlation and dependencies between the configuration parameters with associated system parameters, using an all pair analysis and a Pearson product-moment correlation coefficient (PPMCC), wherein using the all pair analysis comprises:

program instructions to test possible discrete combinations of the configuration parameters and the system parameters, using test vectors to parallelize the tests of parameter pairs;

program instructions to determine dependencies between the configuration parameters with performance metrics; and

program instructions to recommend to a user, configuration settings of the application, based on the analysis of the data.

16. The computer system of claim 15 , further comprising:

program instructions to register for an account for access by the user; and

program instructions to receive access to the configuration settings recommended to a plurality of users of the application running on the plurality of devices.

17. The computer system of claim 15 , further comprising:

subsequent to the recommendation, program instructions to continue to use a current configuration setting; and

program instructions to test the recommended configuration settings in a separate device with workload replication.

18. The computer system of claim 15 , further comprising:

program instructions to automatically change current configuration settings of the user to the recommended configuration settings.

19. The computer system of claim 15 , wherein program instructions to determine correlation and dependencies between the configuration parameters with associated system parameters using the PPMCC comprise:

program instructions to construct a correlation matrix that includes a covariance of two variables divided by a product of a standard deviation of each of the two variables, wherein the two variables are a value for one of the configuration parameters and a value for one of the system parameters.

20. The computer system of claim 15 , wherein program instructions to determine dependencies between the configuration parameters with performance metrics comprise:

program instructions to build a performance model for a given value of one of the system parameters, using a plurality of values for the configuration parameters associated with the given value of one of the system parameters and values for one of the performance metrics by creating a multi-dimensional graph, wherein the configuration parameters, the system parameters, and the performance metrics, each, represent a separate axis on the graph; and

program instructions to determine which value of the plurality of values for the configuration parameters corresponds to one of the values for one of the performance metrics when the value for one of the performance metrics is at a highest point for the given value of one of the system parameters.