Memory analyzer

Modules

• analyzer
• app
• sandbox

Module analyzer

Contains the created dump analyzing library. There's an integration test that expects the sandbox/data/test-heapdump-10.hprof file to exists as a part of the test suite.

Module app

Contains the app that interacts with the user. The app is CLI and contains the following parameters:

• -p / --path - Path to the source HPROF file to analyze (required).
• -l / --list - If specified, the "list namespaces" action will be performed.
• -n / --namespace - If specified, the "waste analysis" action will be performed. As a parameter it expects the namespace to filter. e.g. com.example.
• -e / --exclude - If specified, the "waste analysis" action will be performed. As a parameter it expects the namespace to exclude. e.g. com.example.
• -f / --fields - Add field values to the found instances (works only with -n or -e).
• -v / --verbose - Prints ids of duplicate instances (works only with -n or -e).
• -i / --interactive - Runs interactive shell allowing to query instance ids.
• -c / --csv - Prints the results also to the results.csv file, along with the console output (only works with -n).
• -h / --help - Help.

Module sandbox

Contains source test data. In data, there is several generated HPROF files -- all files beginning with test-heapdump come from the laboratory app, the other one from the Spring Boot app. In example-app, there's a source code for the laboratory app I created for purpose of testing. In spring-boot-example, there's a source code for Spring Boot Hello World app.

Directory test

In this directory, there's the output of the test suite I run for the document. They contain redirected output from the console. They were run via the attached Bash files - test.sh runs the analysis of the test-heapdump-* files (all of them) I mentioned above. The test-spring-boot.sh runs the Spring Boot file analysis (for different namespaces). The test suite is run 3 times for all the commands and the results are given a suffix with the number of run.

Examples

All examples expect the CWD to be the root of the project (memory-analyzer).

Compile

mvn clean package

List the namespaces

java -jar app/target/app-1.0-SNAPSHOT.jar -p sandbox/data/test-heapdump-10.hprof -l

Run the analysis

java -jar app/target/app-1.0-SNAPSHOT.jar -p sandbox/data/test-heapdump-10.hprof -n cz.mxmx.memoryanalyzer.example

Run the analysis and print out the fields

java -jar app/target/app-1.0-SNAPSHOT.jar -p sandbox/data/test-heapdump-10.hprof -n cz.mxmx.memoryanalyzer.example -f

Run the analysis, print out the fields and produce CSV output

java -jar app/target/app-1.0-SNAPSHOT.jar -p sandbox/data/test-heapdump-10.hprof -n cz.mxmx.memoryanalyzer.example -f -c