We would like to announce our 19th Software Performance Meetup. It will take place on May 21th 2019 and will be hosted by codecentric AG.

Jens Wilke, author of cache2k and contributor to the JCache/JSR107 standard, is our first speaker and will talk about  “Java In-Process Caching – Progress, Performance and Pitfalls”. In the second presentation, Fabian Warthenpfuhl und Jonas Kunz of Novatec GmbH will talk about “Improved Observability Using Automated, OpenCensus-based Application Monitoring Solutions”:

Java In-Process Caching – Progress, Performance and Pitfalls

An in-process caching library is part of every Java application stack. Right now EHCache 2 is beeing phased out. The development of Guava Cache is stalling and the “new kids” Caffeine and cache2k enter the stage. The talk will give an orientation about the challenges of high performance caching on modern computer hardware and discuss the different existing technical approaches together with some performance data for comparison.

Improved Observability Using Automated, OpenCensus-based Application Monitoring Solutions

Today’s complexity of enterprise software systems increases the importance of observability and monitoring solutions. DevOps teams or the developers and operators in the “classical world”, respectively, need a clear transparence about the performance, availability and reliability behaviour of their software-systems to be able to manage an appropriate level of service quality. Besides the big, commercial Application Performance Management (APM) solutions the open-source market provides a huge variety of different tools targeting different aspects of APM. Combining these tools opens a great potential for building flexible, tailored APM solutions. Open standards, such as OpenTracing, OpenCensus or OpenMetrics allow for vendor independent collection of data. However, most of the open tools that are based on these standards still require manual adoption of code to instrument applications for data collection. Often, such code changes are not desired or simply just not possible.

In this talk we will present an open, OpenCensus-based approach that provides a simple, yet flexible, configuration-based way of collecting monitoring and business data. Data can be collected without the need for manual code changes. As OpenCensus is used under the hood, this approach allows to create tailored APM solutions, as monitoring backends can be exchanged without any additional overhead.

In this talk, we give insights into the OpenCensus standard and show how OpenCensus instrumentation can be automated. By means of practical examples using different open tools, we demonstrate the flexibility of the presented approach.

Registration