How to approach performance benchmarking of large Java application with JMX?

I'm going to performance benchmark a large Java application. What is the best approach to do it? It is POJO-based application that consists online part with UI and WebService layer, but the most of heavy business logic is in long running batch processes implemented with Spring Batch. The application is deployed on JBoss Application Server and persists data in Oracle database. What I'm looking to get is not only the throughput of batch process or number of concurrent requests handled by online layer. I'm also looking to get more insight and gather some internal metrics for things like batch job steps, JDBC etc. Furthermore I'm going to correlate the metrics with system-wide metrics I can get from JVM, OS, application server, database. Ideally I'd like to do it with no code modifications if possible. My idea to implement this kind of performance monitoring is to rely on JMX and provide a central component that can gather metrics from various MBeans and correlate them. It seems that each main component I want to measure is JMX-enabled: - JBoss Application server has extensive JMX capabilities - Spring Batch can be JMX-enabled with Spring Batch Admin extention - JVM is OK for all it's internals exposed via JMX - Oracle with DMS MEtrics also give good insight into performance of JDBC link I was wondering if there are any good open source frameworks/tools that could help me with collection, correlation and visualization of the metrics as described above. Thanks in advance for your recommendations. P.S. I have done my homework and browsed the open source scene with that regards, but I don'w want to put any specific names here yet to avoid biased answers.

以上就是How to approach performance benchmarking of large Java application with JMX?的详细内容,更多请关注web前端其它相关文章!

赞(0) 打赏
未经允许不得转载:web前端首页 » JavaScript 答疑

评论 抢沙发

  • 昵称 (必填)
  • 邮箱 (必填)
  • 网址

前端开发相关广告投放 更专业 更精准