7–9 November 2016: Conference & Tutorials
9–10 November 2016: Training
Amsterdam, The Netherlands

Ops-friendly microservices: Front to back

Rick Fast (Expedia)
14:00–17:30 Wednesday, 9/11/2016
Average rating: ****.
(4.77, 13 ratings)

Prerequisite knowledge

  • A working knowledge of Docker and HTTP web services

Materials or downloads needed in advance

  • A laptop with Docker and Docker Compose installed (You'll be provided a list of Docker images to pre-pull before the tutorial and given docker-compose files for each part of the tutorial.)

What you'll learn

  • Learn to build a resilient, dynamic, and reliable microservices infrastructure and chassis

Description

Using repeatable/shareable environments built with Docker Compose, Rick Fast demonstrates examples of how to build ops-friendly microservices and the environments they live in, discussing differing options for microservice architectures without being opinionated about specific technologies. Rick offers a high-level overview of basic concerns before deep diving into each. Each section includes runnable code examples using local infrastructure.

Outline:
Building the chassis

  • Common behaviors that need to be built in to make microservices op friendly
  • Ways to bake monitoring, logging, service discovery, and resiliency into libraries shared by all microservices

Service discovery

  • Registering and deregistering microservices for discovery
  • Different mechanisms for finding services in dynamic environments: DNS + server-side load balancing, service discovery + client-side load balancing (using Consul), and service discovery + server-side load balancing (using Consul)

Logging

  • Centralized logging infrastructure
  • Diskless logging with Docker

Monitoring

  • Declarative custom metrics collection from within microservices
  • Making metrics available to monitoring systems (pull method with Prometheus)
  • Pushing metrics (Graphite + Influx)

Resiliency

  • Circuit breakers + timeouts: Insulating your service from upstream failures (with Hystrix)

Building + delivering microservices (with Docker)

  • Building release artifacts as Docker images
  • Deploying Docker images with a scheduler (Mesos + Marathon)
  • Tying it all together with Jenkins pipelines (using Git/gogs, Jenkins 2, and Marathon)
Photo of Rick Fast

Rick Fast

Expedia

Rick Fast is a distinguished software engineer at Expedia. Previously, Rick was an engineer at CDK Global/Cobalt in Portland, an industry leader in automotive retail software, and served as lead engineer at Nokia HERE (formerly Navteq), a provider of digital map data.

Comments on this page are now closed.

Comments

Picture of Rick Fast
Rick Fast
14/11/2016 13:00 CET

I already did. Not sure why they’re not available. Here they are:

https://drive.google.com/open?id=0B4Hx5QV6A8vyQWZMX1VNYVllTWs
Tsvetan Gaydev
14/11/2016 10:57 CET

When do you plan to upload the slides?