ngrinder-3.5.0

3.5.0 (2020.07.31)

• New Feature

  • Introduce server mode agent
    • In previous version, nGrinder agents made TCP connections from agent to controller and this made hard to locate a nGrinder controller behind a firewall. From version 3.5.0, nGrinder introduces server mode agent so that a controller make a connection from a controller to agents.
    • See https://github.com/naver/ngrinder/wiki/server-mode-agent
  • Update agents automatically
    • When the test is executed with older version of agent, it triggers agent auto update.
  • Support GitHub script storage
    • Users can save and run their test script saved in not only embedded SVN server but also external GitHub repository. GitHub enterprise deployed in each own organization is also supported.
    • See https://github.com/naver/ngrinder/wiki/github-script-storage
  • Support MySQL
    • Since 3.5.0, we support H2 and MySQL as databases and fade out Cubrid and provide cubrid-to-mysql migration script.
    • See https://github.com/naver/ngrinder/wiki/mysql-migration
  • Improve test restart time
    • In previous version, nGrinder agent doesn't cache executed test lib & resources, Therefore, it took long time to execute tests in the case of using large amount of libraries and resource. From nGrinder 3.5.0, the test execution is much faster by sending only changed test resources to agents
  • Fade out nGrinder-recorder
    • nGrinder-recorder is no longer supported.
  • Support multiple user defined statistics
    • See https://github.com/naver/ngrinder/wiki/multiple-user-defined-statistic-in-ngrinder
  • Ignore error count
    • Some of users want to continue testing even if there were many errors in the test. So nGrinder 3.5.0 supports that as a test option in advanced configuration panel
  • Support custom classpath
    • If you want to set classpath for test scripts, you can use this.
    • See https://github.com/naver/ngrinder/wiki/custom-classpath

• Modernize

  • Back-end
    • Spring boot
      • Change from spring framework 4 to spring boot 2.
      • Remove embedded Jetty.
    • Hazelcast
      • Use Hazelcast to share data among clusters
      • Reduce load of database.
    • Gradle
      • Use Gradle 6.3 as a build system.
  • Front-end
    • Vue.js
      • Introduce SPA for imporve responsibility & usability.
      • Use webpack as a module bundler.
    • UX/UI
      • Use Billboard.js to draw charts.
      • Bump bootstrap up from 2 to 4.

• Bug fix

  • #490 Fix log file name for agent command line
  • #558 Fix agent state API to work in private agent
  • #571 Prevent zombie agents from remaining in distributed cache
  • #581 Set default locale to "en" when fail to detect user locale
  • #587 Handle case of no sampling model
  • #597 Fix log replacement to hide home path correctly