RUG::B

It's never easy to manage the vast flood of logs that come from a production web application. Tools like Logstash or Graylog offer a way to turn your logs back into useful, searchable data. When I first heard about these tools I wanted to start using them straight away, but there's a catch. Logstash offers powerful regular expressions to help you extract the useful information from your logs, but in practice I found this technique turned into a nightmare very quickly.

Semantic logging means choosing a format that's easy for another program to parse (and hopefully human-readable too); these logs are easy to feed into Logstash or other tools. There's another catch: every library, gem, project, code snippet that logs information has its own format already. Is it possible to take control of this flood of information, get everything into a search engine and never again have to use the humble grep command to search your logs?

Using a bit of Ruby magic I believe it's possible and even fairly simple to implement. In this talk I explain the problems I faced and why I think you should consider adopting semantic logging in your project as well.

InfluxDB is a really good Time Series Database with a nice tooling around it called the tick stack.

In this talk I will give you a short introduction to it. What is it? What tools does it offer? For what can I use it?

Everybody is talking about microservices, but how do you go about actually building one and how do you communicate with it? Sneakers is a nice Ruby library on top of RabbitMQ that makes this very easy - if you know Sidekiq you'll feel right at home. I'll describe how to build a microservice with Sneakers, how to daemonise it and have proper logging and how to communicate with it from some other project.