At the 3rd ACM/SPEC International Conference on Performance Engineering (ICPE 2012) in Boston (MA, USA) we gave an invited poster/demo presentation of our Kieker framework.

• Paper: ( .pdf | bib )
• Poster: ( .pdf )
• Presentation: ( .pdf )

Today, we released version 1.5 of our Kieker framework for application performance monitoring and dynamic software analysis.

Major new features:

• Completely reworked pipe-and-filter framework in Kieker.Analysis
  • Use of Java annotations to specify plugins, repositories,
    input/output ports etc.
  • Support for multiple readers
  • Load and save configurations from/to .kax files
    (based on a new meta-model specified using Ecore)
  • Basic tools to execute/visualize .kax files
• New database writer/reader supporting arbitrary record types
• Support for immutable record types
• New binary filesystem writer/reader
• New print stream writer
• New re-usable filters (tee, counting, timestamp etc.)
• Complete method signatures used in probes/records for control-flow tracing
• New record types/probes for event-based tracing
  (method call, entry, exit etc.)
• Re-enabled Java 5 support
• Removed dependency to commons.logging
• Additional Kieker Jar usable as javaagent for AspectJ-based instrumentation

Quality:

• Bug fixes
• Added a number of (JUnit) tests

See the release notes for details.