Optimization analysis using similar frequencies
View Patent ↗Tracer objectives in a distributed tracing system may be compared to identify input parameters that may have a high statistical relevancy. An iterative process may traverse multiple input objects by comparing results of multiple tracer objectives and scoring possible input objects as being possibly statistically relevant. With each iteration, statistically irrelevant input objects may be discarded from a tracer objective and other potentially relevant objects may be added. The iterative process may converge on a set of statistically relevant input objects for a given measured value without a priori knowledge of an application being traced.
1. A method performed by a computer processor, said method comprising:
receiving an application to instrument;
identifying a plurality of trace objectives for said application, said trace objectives comprising trace objects and input parameters;
causing said plurality of trace objectives to be executed with said application to generate a plurality of results sets;
selecting a first results set generated from a first tracer objective;
analyzing said first results set to identify a first frequency;
identifying a second trace objective having a second results set comprising said first frequency;
identifying a second input parameter in said second trace objective;
creating a third trace objective comprising said first trace objective and said second input parameter; and
causing said third trace objective to be executed.
2. The method of claim 1 further comprising:
analyzing said first results set to create a first frequency signature, said first frequency signature comprising said first frequency.
3. The method of claim 2 further comprising:
searching said plurality of trace objectives to identify said second trace objective having a second frequency signature matching said first frequency signature.
4. The method of claim 3 , said first frequency signature and said second frequency signature being matched using a similarity analysis.
5. The method of claim 4 further comprising:
determining a similarity score for each of said plurality of trace objectives;
said similarity score representing a similarity strength between said first results set and results sets for each of said trace objectives.
6. The method of claim 2 , said frequency signature comprising a plurality of frequencies.
7. The method of claim 6 , said first frequency being a dominant frequency.
8. The method of claim 1 further comprising:
determining that said second input parameter has a statistically relevant correlation to a second observed parameter data.
9. The method of claim 1 further comprising:
determining that a first input parameter does not have a statistically significant relationship to a first set of observed parameter data; and
removing said first input parameter from said third tracer objective.
10. The method of claim 9 further comprising:
analyzing a third set of observed parameter data having been generated by said third tracer objective;
determining that said second input parameter does not have a statistically significant relationship to said third observed parameter data; and
removing said second input parameter from said third tracer objective.
11. The method of claim 1 , said first frequency being identified using autocorrelation analysis.
12. The method of claim 1 , said results sets comprising time series data.
13. A system comprising:
a processor;
a communications link to a dispatcher, said dispatcher that:
identifies a plurality of trace objectives for an application, said trace objectives comprising trace objects and input parameters;
causes said plurality of trace objectives to be executed with said application to generate a plurality of results sets;
an analyzer executing on said processor that:
selects a first results set generated from a first tracer objective;
analyzes said first results set to identify a first frequency;
identifies a second trace objective having a second results set comprising said first frequency;
identifies a second input parameter in said second trace objective;
creates a third trace objective comprising said first trace objective and said second input parameter; and
causes said third trace objective to be executed by said dispatcher.
14. The system of claim 13 , said analyzer that further:
creates a first frequency signature from said first results set, said first frequency signature comprising a first plurality of frequencies; and
creates a second frequency signature from said second results set, said second frequency signature comprising a second plurality of frequencies.
15. The system of claim 14 , said second trace objective being selected by comparing similarities of said first frequency signature and said second frequency signature.
16. The system of claim 13 , said analyzer that further:
identifies a first input parameter in said first tracer objective;
determines that said first input parameter has low statistical significance in relation to said first results set; and
removes said first input parameter from said third trace objective.
17. The system of claim 16 , said analyzer that further:
creates a trace model comprising a first set of input parameters and said first results set.
18. The system of claim 17 , said trace model being created using a curve fitting technique.
19. The system of claim 13 , said dispatcher executing on said processor.