Techniques for workload toxic mapping

1. A method, comprising:

obtaining, by a processor, metrics for a workload, resources of the workload, a processing environment that processes the workload, and other resources of the processing environment by sampling micro performance values in microsecond intervals of different combinations of the workload, the resources, the processing environment, and the other resources, and wherein the workload is a Virtual Machine (VM);

wherein obtaining further includes obtain at least some of the metrics as a time for response made by an external service when the external service is presented with a request, and wherein the request is at least part of the workload;

storing, by the processor, the metrics in a toxic map;

mapping the micro measurements in the toxic map, wherein the toxic map is represented as a data structure associated with the workload, a workload state, the processing environment, a processing environment state, the resources, and the different combinations;

generating an index value for each workload, workload state, processing environment, processing environment state, resource, and each of the different combinations used;

mining, by the processor, the toxic map using the index values and identifying in response to the mining at least one performance problem with one or more of: the workload, the resources, the processing environment, the other resources, and one or more of the different combinations;

producing a baseline for results based on evaluation of the micro performance values; and

comparing the results to previous baselined results using policy and take one or more automated actions;

processing at least one of the automated actions using a load balancer to execute the workload on a different machine from a machine associated with the workload.

2. The method of claim 1 , wherein obtaining further includes obtaining measurements reflecting at least some of the metrics at predefined intervals.

3. The method of claim 1 , wherein obtaining further includes obtaining the workload state for the workload reflecting at least one of the metrics.

4. The method of claim 1 , wherein obtaining further includes obtaining a status for the processing environment reflecting at least one of the metrics.

5. The method of claim 1 , wherein mapping further includes mapping each metric to a unique resource associated with that metric.

6. The method of claim 5 , wherein mapping further includes mapping each metric to a unique resource state for that unique resource.

7. The method of claim 1 , wherein mining further includes evaluating a policy and processing decisions rules when mining the metrics to identify conditions mapped to the at least one performance problem.

8. A method, comprising:

evaluating, by a processor, metrics captured for: a workload, a processing environment that processes the workload, resources of the workload, and other resources of the processing environment over a network for identifying relationships between: the workload, the processing environment, the resources, and the other resources, wherein evaluating further includes sampling micro values at microsecond intervals for combinations of: the workload, the processing environment, the resources, and the other resources, and wherein the workload is an Operating System (OS) that processes in the processing environment, wherein evaluating further includes obtaining at least some of the metrics as a time for response made by an external service when the external service is presented with a request, and wherein the request is at least part of the workload, wherein evaluating further includes producing a baseline for results associated with evaluation of the metrics;

mapping, by the processor, the micro values in a toxic map that is represented by a data structure associated with the workload, a workload state, the processing environment, a processing environment state, the resources, and the combinations;

generating, by the processing, an index value for each workload, workload state, processing environment, processing environment state, resource, and each of the combinations;

processing, by the processor, decisions based on a policy applied during the evaluating for identifying a performance scenario deemed problematic with one or more of: the workload, the processing environment, the resources, the other resources, and specific ones of the combinations by mining the toxic map using the index values;

providing, by the processor, particular metrics relevant to the performance scenario to a principal operating within the workload;

comparing the results to previous baselined results using a second policy; and

performing at least one action from a plurality of automated actions that improves the performance scenario, and processing a particular one of the plurality of automated actions using a load balancer that executes the workload on a different processing environment.

9. The method of claim 8 , wherein evaluating further includes weighting the metrics in response to rules processed during the evaluation.

10. The method of claim 8 , wherein evaluating further includes capturing the metrics at regular intervals of time.

11. The method of claim 8 , wherein evaluating further includes mapping each metric to a particular resource of the workload.

12. The method of claim 8 , wherein evaluating further includes mapping at least some metrics to resources associated with the processing environment and unrelated to the workload.

13. The method of claim 8 , wherein providing further includes packaging the metrics relevant to the performance scenario in a principal-defined format for delivery to the principal.

14. A system, comprising:

a processor; and

a mapper configured to:

i) execute on the processor, ii) take metrics at predefined microsecond intervals from: a) a workload represented by a Virtual Machine (VM) or Operating System (OS) that processes in a processing environment over a network, b) the processing environment, c) resources of the workload, and d) other resources of the processing environment by sampling, at a micro second level, performance values from the metrics for combinations of: the workload, the processing environment, the resources, and the other resources, iii) obtain at least some of the metrics as a time for response made by an external service when the external service is presented with a request and the request is at least part of the workload, and produce a baseline for results based on an evaluation of the performance values, iv) store the metrics in a toxic map, v) map the performance values in the toxic map, wherein the toxic map is a data structure associated with the workload, a workload state, the processing environment, a processing environment state, the resources, and the combinations, vi) generate an index value for each workload, workload state, processing environment, processing environment state, resource, and each of the combinations, and vii) process a policy and decision rules to mine the toxic map using the index values and the toxic map to identify a performance scenario with one or more of: the workload, the processing environment, the resources, and the other resources, and viii) compare the results to previous baselined results using a second policy and process one of a plurality of automated actions to improve the performance scenario, and process at least one of the plurality of automated actions using a load balancer that executes the workload on a different processing environment.

15. The system of claim 14 , wherein the automated action is a notification sent to a principal regarding the performance scenario and those metrics determined to be relevant to the performance scenario.