search

lcabaceira / crystal

Alfresco predictions, analisys, benchmarks
3 stars 1 forks source link

Analise CSV Files structure, defining dimensions and measures model #2

Closed lcabaceira closed 8 years ago

lcabaceira commented 8 years ago

Francesco to analise CSV files structure and provide feedback about enteties , model, structure of data-mart

lcabaceira commented 8 years ago

francesco_csv.zip

fcorti commented 8 years ago

I have found a good starting point. http://alexandru-ersenie.com/2011/05/12/java-ee-load-performance-testing-and-reporting-using-open-source-tools-part-1-concept-examples-and-sample-reports/

We could be inspired for the first part of the project (the data structure and etl). I don't know why but the author decided to use Jasper in a Pentaho environment... I would prefer to remain on the Pentaho path.

Note 2015-Dec-20: it's a custom import. Prefer to develop it from scratch.

lcabaceira commented 8 years ago

Yes, that scripting part is really necessary. That was the part i was mentioning yesterday, where we use remote scripting (mainly .sh or pearl scripts) to obtain infra-structure metrics during the tests execution. Those should also generate .csv files to feed the pentaho data extractor.

fcorti commented 8 years ago

Hi all,

Things are starting to become more clear.

According to the documentation page (http://wiki.apache.org/jmeter/LogAnalysis) all the report plugins should generate the same data structure (I'm quite sure but more checks will be done later). For this reason we could base all the data migration to Pentaho on a single report extraction for each test (for example using a Simple Data Writer plugin). Of course the users will be free to use a different plugin for interaction with Pentaho, or even define multiple plugins for her needs.

Saying that, the reasonable constraints I'm going to define are listed below:

${__BeanShell(import org.apache.jmeter.services.FileServer; FileServer.getFileServer().getBaseDir();)}${__BeanShell(File.separator,)}..${__BeanShell(File.separator,)}JMeterLogs${__BeanShell(File.separator,)}${__TestPlanName}_${__time(yyyy-MM-dd_HH-mm-ss,)}_${__UUID}.xml

This string will instruct JMeter to save the report into the <crystal>/TestCases/JMeterLogs folder with the name <ScriptName>_<date/time>_<UUID>.xml.

Please share comments and suggestions if something should be defined differently.

Added this page tot he wiki: https://github.com/lcabaceira/crystal/wiki/Developing-your-test-case

lcabaceira commented 8 years ago

Looks good, i will change my cmis test to reflect that.