Track: Production Readiness: Building Resilient Systems

Location: Ballroom A

Day of week:

A production readiness review is used by software companies to determine whether the design and implementation of the system is ready to be released to its customers. The process is used to identify and address the reliability of a service, sufficiency of the coverage of privacy and security needs, and the ease of the operability. This track explores what types of aspects of software need to be prepared to start taking on full production load with customer’s data. Topics include observability, emergency response, capacity planning, release processes, and SLOs for availability and latency.

Track Host: Aysylu Greenberg

Senior Software Engineer @Google

Aysylu Greenberg is a Sr Software Engineer at Google working on infrastructure and the Eng Lead of the Grafeas and Kritis open source projects. In her spare time, she ponders the design of systems that deal with inaccuracies, enthusiastically reads CS research papers, and dances.

10:35am - 11:25am

Capacity Planning for Crypto Mania

Over the course of 2017, Coinbase experienced exponential user and trading volume growth, which in turn led to periods of website instability and downtime. During this period, we saw our systems perform at the very edge of their capacity which inspired important capacity and performance improvements. Since then we have sought new ways to push our systems to their limits so that we can be sure that we are focusing our energy on the right projects. 

Come to hear how Coinbase engineers are applying lessons from these experiences to create new tools and techniques for capacity planning to prepare for future waves of cryptocurrency enthusiasm.

Jordan Sitkin, Software Engineer @coinbase
Luke Demi, Software Engineer @coinbase

11:50am - 12:40pm

Building Production-Ready Applications

In 2016, Susan Fowler released the 'Production Ready Microservices' book. This book sets an industry benchmark on explaining how microservices should be conceived, all the way through to documentation. So how does this translate actionable items? This session will explore how to expertly deploy your microservice to production. The audience will learn best practice for designing, deploying, monitoring & documenting application. By the end of the session, attendees should feel confident that they have the knowledge to deploy a service that will be reliable and scalable.

Michael Kehoe, Architect of reliable, scalable infrastructure @LinkedIn

1:40pm - 2:30pm

Building Resilience in Production Migrations

How do you migrate stateful systems with confidence? Especially when downtime is not an option? Netflix Billing Infrastructure needs to be up 24/7 to support 130+ million global customers. Billing services are the source of truth for a customer’s billing state which changes as customers apply gift cards, update their Method of Payment or are just charged every month. We want this experience to be seamless and accurate anytime, day or night, so our databases are constantly in action, with no possible downtime windows.   

We have succeeded in multiple major rewrites. In the first one, we migrated billions of rows from Oracle in our data center to MySQL AWS Cloud. Post that, we also rewrote Netflix Balance Service that manages gifts and promotions redeemed by customers. In the recent one, we rewrote our legacy invoice processing system and seamlessly transitioned from a MySQL solution to Cassandra. All these efforts involved different strategies that helped us achieve these with a flip of a switch and without anyone realizing that Netflix had been doing major overhauls of its Billing Infrastructure. We will share our migration stories and what helped us build resilience.

From this talk, attendees will learn:

  • Baseline Considerations to be thought through in migrations.
  • Variations and aspects of achieving zero downtime especially around state migration.
  • Tools and technologies that can be helpful.

Sangeeta Handa, Engineering Manager, Billing Infrastructure @Netflix

2:55pm - 3:45pm

CRDTs in Production

In search of scalability and availability improvements, many companies adopt eventual consistency as the consistency model underlying their stateful systems and persistent data stores. At the same time, software designers are focused on creating resilient systems ready to work in production with minimal complexity. Dmitry will share lessons learned in developing a distributed system based on an eventually consistent data store. The final solution utilizes conflict-free, replicated data types with causality tracking to achieve strong eventual consistency for critical data in multi-master, multi-datacenter DB (Aerospike) deployments.

Dmitry Martyanov, Software Engineer @PayPal

4:10pm - 5:00pm

Yes, I Test In Production (And So Do You)

Testing in production has gotten a bad rap.  People seem to assume that you can only test before production *or* in production.  But while you can rule some things out before shipping, those are typically the easy things, the known unknowns.  For any real unknown-unknown, you're gonna be Testing in Production.
And this is a good thing!  Staging areas are a black hole for engineering cycles.  Most interesting problems will only manifest under real workloads, on real data, with real users and real concurrency.  So you should spend much fewer of your scarce engineering cycles poring over staging, and much more of them building guard rails for prod -- ways to ship code safely, absorb failures without impacting users, test various kinds of failures in prod, detect them, and roll back. 
You can never trust a green diff until it's baked in prod.  So let's talk about the tools and principles of canarying software and gaining confidence in a build.  What types of problems are worth setting up a capture/replay apparatus to boost your confidence?  How many versions should you be able to run simultaneously in flight?  And finally we'll talk about the missing link that makes all these things possible: instrumentation and observability for complex systems.

Charity Majors, Co-Founder @Honeycombio, formerly DevOps @ParseIT/@Facebook

Last Year's Tracks

This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.