Probabilistic scripts for automating common-sense tasks
If you work with BigData, you may need to clean data. This talk shows a way through probabilities. By applying Bayes’ theorem, one can deduce missing values or incorrect values. More and more, probabilistic approaches give a good mean to reduce hard work and that is quite neat.
Linux Pressure Stall Information (PSI) by Example
when looking at linux metrics, load average gives some infos on how your server handles work. But, it mixes CPU, IO and RAM. New kernels exposes more precise metrics with PSI. This article goes in deep on what it means and how to use it.
Linus Torvalds isn’t worried about Microsoft taking over Linux
For some time now, Microsoft has been embracing open source technologies such as GNU/Linux. So much that it’s the most popular deployed operating system. Linus Torvalds gave in this article his thoughts about the adoption of Linux kernel by Microsoft.
New french administration source code portal (FR)
In the last few years, french administration trends to liberate some of his source code and data. In october, a brand new source code portal has been released. It aims to centralize all the source code published by french administration offices.
How LinkedIn customizes Apache Kafka ?
How to securize a REST API (FR)
API securization could be complicated to implement. There a plenty of standards and ways to do it. OCTO published an summary of all the existing solutions and weight pros and cons for each of them.
Terraform or CloudFormation
Infrastructure as Code is the way to go to define and easily modifiy your infrastructure (see also this article from Loïc Rodier on that topic). Several tools exist, the main one is Terraform. A nice feedback about CloudFormation (from AWS) comparing to Terraform.
Performance and diagnosis, methods and tools (FR)
When doing a performance analysis or understanding an incident, lots of commands are done on the system to get some insights of what could be wrong. This slideck smashes them all plus give methods to apply them: what to analyze first with what command, what does the result mean. A complete documentation to keep close by. Linux only of course ;)