High Availability at Braintree

Explore high availability strategies at Braintree, a payment gateway, in this 36-minute conference talk from Strange Loop 2013. Discover how to minimize downtime and maintain service reliability in a Ruby-based application. Learn techniques for managing planned maintenance, including traffic pausing, rolling deploys, and schema changes without interruptions. Understand approaches to mitigate unplanned failures through load balancing, redundant infrastructure, and application-level retry mechanisms. Gain insights into Braintree's refined deployment process, stateful services management, and networking solutions. Examine the importance of safeguards, monitoring, and automation in reducing errors and maintaining system integrity. Benefit from the speaker's extensive experience in software development and payment processing infrastructure.

Intro
Ruby on Rails
Refined deploy process
Rails caches columns
We run multiple versions of the code at once
Limitations
Lightweight rack adapter
Build our own
Stateful services
Networking - inbound
Processor proxies
Connection failures
Summary - unplanned failures
People make mistakes
Reduces errors
Safeguards
Monitoring Nagios
Summary - people Automation