2016 October

#

25 Oct: Efficient logging with Spring Boot, Logback and Logstash

Logging is an important part of any entreprise application and Logback makes an excellent choice: it’s simple, fast, light and very powerful. Spring Boot has a great support for Logback and provides lot of features to configure it. In this article I will present you an integration for an entreprise logging stack using Logback, Spring Boot and Logstash. WARNING The Spring Boot recommands to use the -spring variants for your logging configuration (for example logback-spring.xml rather than…

#

17 Oct: Advanced tools: playing with Java Native Access

This post results from a recent deep diving in the source code of Elasticsearch, which uses JNA mainly for memory management when configuring the mlockall. I will present you how to use JNA in a very simple example: how to check the user who has launched the JVM. Note Remember that when you are thinking about solution using OS native calls, you must deal and depend with platform librairies. So use it carefully and only for specific…

#

10 Oct: OS monitoring with… Java

Sometimes it may be useful to get system information like the usage of a disk or the available network interfaces. For instance, Elasticsearch use this kind of tools in order to display at startup time some infos about open file descriptors or the size of the direct memory available for the JVM. The aim is not to replace a real system monitoring agent, but to guide the user to take advantage of the product by configuring it…

#

06 Oct: Find and kill slow running queries in MongoDB

In Mongo, or more generally in any data storage engine, queries or updates that take longer than expected to run can be caused by many reasons: – Slow network – Wrong schema design (we all have seen the famous all-in-one table…) – Wrong database design (“let’s store 100To of data in a standalone mongod!”) – Bad partitioning (Hbase table with 200 regions with 2MB of data) – Lack of useful indexes – No statistics – Incorrect hardware…