research-iobserve / iobserve-analysis

Tools, probes and libraries used in iObserve to monitor and analyze software, as well as to plan and execute its modification (MAPE-K loop)

iObserve Monitoring and Analysis

This repository contains the sources for the iobserve analysis including

• common for the monitoring records (event types)
• monitoring for the monitoring probes
• analysis for the analysis plugins on basis of teetime
• analysis.cli command line version of the analysis
• analysis.service the service version of the analysis
• splitter separates monitoring data streams by host
• rac.creator this tool allows to create a RAC mapping based on monitoring data, a PCM model and a supplementary mapping file.
• test.setup test tooling

The code is in a prototype stage and not fully functional, especially, as some parts are still located in other SVN repositories.

Setting up iObserve gradle build

Prerequisites

• We use Java 8. you need an Java 8 JDK installed on your machine.
• You need additional packages which are compiled in iobserve-repository. Checkout the iobserve-repository git repository alongside iobserve-analysis git clone [e-mail]:research-iobserve/iobserve-repository.git
• You need Kieker example data, a RAC and a corresponding palladio model

'Note: This is outdated, we switched to gradle'

Setup Build Environment

In the iobserve-analysis directory, create a file gradle.properties. Add the following variable declaration to that file and save it.

api.baseline=PATH_TO_IOBSERVE_REPOSITORY/mvn-repo/

Substitute PATH_TO_IOBSERVE_REPOSITORY with the actual absolute path on your computer to the iobserve-repository directory.

Compile

Depending on your local setup you may use:

• ./gradlew build (linux, mac, etc.)
• gradlew.bat build (windows)
• gradle build (in case you have gradle 3.2.1 installed on your system)

Execution

After compilation, you may find in various folders packages containing executables, including the analysis, the replayer, the session reconstructor, etc. The process to use them is quite similar. In the following we introduce the most common tools.

Alongside all executables we have example configurations placed in the respective sub project folder with the suffix config. All executables support an environment variable project_OPTS where project must be substituted by the executable name. for analysis it is ANALYSIS_OPTS. Through this mechanisms further parameter can be specified which is quite helpful.

To activate logging use: export ANALYSIS_OPTS="-Dlog4j.configuration=file:///full/qualified/pathname/log4j.cfg" Please note that you have to adjust path and file name to your setup.

Analysis

The analysis is located in analysis/build/distributions/ You may find two archives containing executables.

• analysis-0.0.3-SNAPSHOT.tar
• analysis-0.0.3-SNAPSHOT.zip Choose your preferred kind of archive and extract it in our analysis directory (preferably outside of the directory of the git repository). After extraction you can start the analysis with:
• analysis.cli-0.0.3-SNAPSHOT/bin/analysis (linux, mac, etc.)
• analysis.cli-0.0.3-SNAPSHOT/bin/analysis.bat (windows)

usage: iobserve-analysis -h,--help show usage information -c configuration file

For a minimal analysis, type: analysis-0.0.3-SNAPSHOT/bin/analysis -c analysis.config

Collector

Usage:

[options] Options:

• -c, --configuration Configuration file.

Log-Replayer

Session Reconstructor

Service Privacy Violation