Getting To, and Through, Our First Black Friday with Critical Apps [I]
summary
Luiza Labs moves Magazine Luiza’s critical apps onto Kubernetes in production since May 2016, builds an open source Heroku-like deployer (Teresa CLI), and hits 700+ production deployments before Black Friday. Clusters run on AWS (mobile site tested on GCP for three months), nodes scale with ASGs and kubectl drain, and ELB misconfigurations force operational fixes. A fallback cluster with weighted DNS sits ready; result is zero Black Friday downtime after prior 60 and 26 minute outages. Go-swagger and single-engine ownership cause pain.
AI-written summary · about
chapters (13)
- 0:00Intro
- 0:19About Arnaldo
- 2:31What is Luiza Labs
- 3:54Why Luiza
- 5:03Black Friday 2014
- 17:26Heroku
- 19:48Issues with Kubernetes
- 23:11Dont touch the environment
- 26:03Focus on Kubernetes
- 28:45Focus on Small Things
- 29:01Mistakes on Tourism
- 29:33Take Away
- 30:54Questions
related talks
53mTwo Black Fridays later: hits and misses
27mRunning Kubernetes in Production A Million Ways to Crash Your Cluster
46mManaging a Billion Transactions with Tech, Tactics, and Teamwork
29mBlack Friday - Lessons in Resiliency and Incident Response at Shopify
24m"Misadventures in Continuous Delivery"
46mKeep Building Fresh: Shopify's Journey to Kubernetes
40mThe Good, the Bad, and the Ugly of Managing Production Systems with Kubernetes
27mA Safari of Kubernetes and its Natural Habitat