The document outlines the concept of microservices as small, autonomous software components that address specific tasks while being loosely coupled and independently scalable. It discusses the conditions under which microservices should be considered, the appropriate size for a microservice, communication methods, and management strategies, including deployment tools and self-healing practices. The complexities and challenges associated with microservices, particularly in terms of infrastructure and dependency management, are also highlighted.

1. Microservices
Pros & Cons
Things to Think About

2. What is a microservice?
 Small piece of software that does one thing really well
 Loosely coupled
 Separate data store
 Just enough to solve a problem
 Right technology for the job
 Autonomous
 Can update as often as is needed
 Intelligence in the service, not the routing/infrastructure/bus
 Immutable infrastructure

3. Should I use microservices?
 It depends!
 Likely no
 Think about fallacies of distributed computing:
 The network is reliable
 Latency is zero
 Bandwidth is infinite
 The network is secure
 Topology doesn’t change
 There is one administrator
 Transport cost is zero
 The network is homogeneous

4. When should I consider microservices?
 Many teams work on the same code base
 Merge hell, cross team dependencies, ...
 Huge monolithic application which is difficult to deploy
 Monolith cannot be scaled horizontally
 Different parts of the application have totally different requirements
 CPU bound
 I/O bound
 Memory bound
 etc.
 Some but not all areas of the application change frequently
 Development stack is outdated. New tools and patterns are hard if
not impossible to embrace

5. How big should a microservice be?
 Small enough to fit full context in your head
 Big enough to solve a problem
 Owned by one team

6. What does “big enough” look like...
Some examples:
 Lambda: https://github.com/meconlin/lambda-generic-microservice
 C#: https://github.com/AFASSoftware/CQRS-Microservices
 Spring: https://github.com/kbastani/spring-cloud-microservice-example
 akka: https://github.com/theiterators/akka-http-microservice
 https://github.com/dustinbarnes/microservice-example

7. How should a microservice communicate?
 Synchronous
 Asynchronous
 Messaging
 Fan out
 HTTP/REST
 TCP/IP
 Pub/sub
...but zero logic in the communication pipeline!

8. Are microservices less complex?
 Code wise, yes
 Deployment wise, yes
 Infrastructure
configuration wise, no
 Dependency management
wise, no

9. Microservice Management
Overview

10. How to manage microservices?
It’s a big world with lots of cutely named tools!
• Deploy: Jenkins / TeamCity / Ansible / Chef / Capistrano / StackStorm
• Discovery/config: Consul / Consul-Templates / etcd / Registrator / skydns
• Containers: Docker / Compose / Vagrant / Otto / Lambda
• Container Clustering: ECS / Kubernetes / Mesos / Docker Swarm
• Request Routing: Nginx / HAProxy / Kong / API Gateway
• Self Healing: Consul / ZooKeeper / Serf
• System Health: hystrix, SumoLogic, Nagios, NewRelic, statsd, LogEntries
AUTOMATE EVERYTHING

11. Deployment
 Continuous Deployment
 Continuous Delivery
 Versioned
 Blue / Green
 A-B Testing
 Net new, add to routing
 Zero downtime
Tools: Jenkins / TeamCity / Ansible / Chef / Capistrano / StackStorm
NEVER DESTRUCTIVE

12. Service Discovery & Configuration
 Minimize known dependencies
 No bottlenecks due to outage allowed
 Down stream health monitoring
 Configuration at deployment time when possible
 Configuration in runtime if you have to (adds fragility/degradation)
Tools: Consul / Consul-Templates / etcd / Registrator / skydns
NO DEPENDENCY KNOWLEDGE

13. Containers & Clustering
 Removes “works on my box” story
 Installed software dependencies become constrained to your need
 No more “servers as pets”
 Infrastructure as code becomes a reality
 Can deploy to fabric/cluster for better auto scaling story
 Serverless truly abstracts hardware from application
Tools: Docker / Compose / Vagrant / Otto / Lambda / AWS ECS /
Kubernetes / Mesos / Docker Swarm
NO HARDWARE DEPENDENCY PREFERRED

14. Request Routing
 Public abstraction from internal details
 Internal location can become dynamic
 Multiple versions of the same thing can be long lived
 Makes deployment story more flexible
 Live traffic can be drained over
 Warming up new instances is possible
Tools: Nginx / HAProxy / Kong / API Gateway
NEVER EXPOSE YOUR SERVICES DIRECTLY

15. Self Healing
It’s not IF it will fail but WHEN it will fail!
 Auto healing
 Automated Remediation
 Circuit breaker
 Fallbacks
 Graceful degradation
 Don’t cascade failures
Tools: Consul / ZooKeeper / Serf
PLAN FOR FAILURE FIRST

16. System Health
 Measure anything, measure everything
 https://codeascraft.com/2011/02/15/measure-anything-measure-everything/
 Performance monitoring, exception monitoring, logs, metrics
 NewRelic, nagios, SumoLogic
 statsd / graphite (hosted graphite) / kibana / grafana
 Centralized logging (logentries, logstash)
 Circuit Breaker (hystrix)
Tools: hystrix, SumoLogic, Nagios, NewRelic, statsd, LogEntries
VISUALIZE EVERYTHING

17. Circuit Breakers with Hystrix

20. Questions?
James Allen
http://www.linkedin.com/in/jamesallenatx
Miguel Gonzalez
https://www.linkedin.com/in/magonz
@doesnotcompile
Gabriel Schenker
https://www.linkedin.com/in/gabrielschenker
@gnschenker
Andrew Siemer
https://www.linkedin.com/in/andrewsiemer
@asiemer