I love giving talks and sharing knowledge, whether by talking about establishing SRE practices or about deep technical problems. The range of topic varies from capacity planning, on-boarding teams to the on-call rotation, ChatOps, to building developer tools for large-scale platforms.
We all love our Sidekiq’s and our Resque’s. But they do let us down sometimes. Not because they are bad, but because the queueing theory is limiting us. There is a way to break out of the madness though - let’s explore how to get our job queues under control.
We’re all used to configure servers with Chef and deploying with Capistrano, however moving to cloud and Kubernetes requires a change in the approach to how we deploy apps. The talk covers common problems that engineers face when setting up Rails apps in Kubernetes.
In this talk I demonstrate what allowed my team at Shopify to run background jobs at scale. It largely focuses on interruptibility long-running work and the new API to describe jobs that we’ve developed to solve the problem.
Monoliths are often considered as a bad thing that is slow to get working locally, run tests and deploy. The talk attempts to break that myth by demonstrating how tou can invest into the monolith and make it a great experience for developers, from the moment when they clone the repo and run tests, to the point when they ship the code.
In this talk I describe how we apply ChatOps at Shopify and how we’ve been able to switch to the commands API without regular expressions, which makes it a lot easier for developers to add new features.
Talking at RailsConf about benchmarking Rails as a framework and RubyBench, a community project for long running benchmarks, which helps to find performance regressions in Ruby and Ruby on Rails.
Presenting Capistrano 3, a major update to the most popular deployment tool for Ruby apps.