Mission and Activities

The RG DevOps Performance Working Group is a forum for individuals and organizations interested in the interplay of DevOps and performance engineering. The mission of the working group is to consolidate concepts and tools to better integrate these activities.

General Topics of Interest

Topics of Interest include, but are not limited to, the following:

  • Performance and workload model extraction
    • Dynamic/adaptive instrumentation techniques
    • Combination of static and dynamic analysis
  • Data feedback between development (Dev) and operations (Ops)
    • Interchange formats to exchange metrics and models
    • Process models to better integrate SPE and APM activities
  • Runtime performance management techniques
    • Reactive/proactive problem detection, diagnosis, and resolution
  • Load testing
  • Benchmarks for SPE and APM methods, techniques, and tools

Activities

  • Landscaping of performance model extraction approaches
  • Knowledge transfer between participating organizations
    • Establish and document best practices
  • Developing new methods, techniques, and tools for SPE and APM integration
  • Common case studies to compare approaches for:
    • Performance and workload model extraction
    • Runtime performance management
    • Load testing

Group activities

In addition to activities involving the group as a whole, the group is organized into subgroups consisting of 6-8 participants who collaborate closely on concrete topics. Currently, the following subgroups are active:

 

1. Performance regression testing of microservices: This subgroup focuses on the challenges of performance regression testing microservices, such as the inevitable differences in microservice deployments which make it hard to compare their performance measurements. The subgroup is currently running experiments on the TeaStore application to demonstrate the impact of these challenges.

2. Model extraction and refinement in continuous software engineering: This subgroup focuses on how performance models and model-based prediction can be leveraged to improve the continuous software engineering process.

3. Performance of continuous delivery infrastructures. This subgroup focuses on the evaluation and improvement of continuous delivery (CD) infrastructures, which have become a critical component of software development. The subgroup is currently analyzing empirical performance data of a CD system. Next steps include the performance modeling of these kinds of systems.

 

The group is also active in organizing international community events

such as the International Workshop on Quality-Aware DevOps (QUDOS) or

Dagstuhl seminars (e.g., https://www.dagstuhl.de/16394 and

https://www.dagstuhl.de/18283).