Green / Blue Deployment with Immutable Servers

Post on 26-Jan-2015

109 views 2 download

description

Talk about green / blue deployments at the University of Applied Sciences Salzburg

Transcript of Green / Blue Deployment with Immutable Servers

Green / Blue Deployment with Immutable Servers

Highly automated deployments

Who am I?

Simon Dittlmannat tado° since 2013previous Payback, WISI, Mivitec...

Developer, DevOpPreferred Language: JS

Contact SimonDittlmann github.com/Pindar

What’s tado°?

The app that controls your heating and saves energy.

With a Connector kit for your home.

What’s tado°?

Introducing tado° Cooling intelligent A/C control

Regular deployments

Assumption: servers are always up and running

but suddenly…

now everything has to be as fast as possible and it starts to be chaotic

Green / Blue Deployment with Immutable Servers● enables you to easily deploy infrastructure● automates and structures the deployment● makes it easy to replace hundreds of servers

within minutes

how?

Running stack

Create new stack

Swapcan also be a soft transition phase

Remove previous stack

Cloudformation

What about log data?

shipper(instance)

shipper (instance)

shipper(instance)

Queue(Redis)

indexing (elastic search)

processing (logstash)

kibana (JS-view)

Backup S3

Cluster

Logstash

Logstash

What about monitoring?

AWS CloudwatchInstances shipping of metric

data(app layer, e.g. heap space)

monitor os level: cpu, network traffic etc

Dashboard

eMail

HTTP call

CloudWatch

self healing

assumption now: servers have problems but you can handle them

Three independent work streams

Application

Infrastructure

Virtual Machine Deployment AWS

Grails, Java

Puppet

Cloudformation

deploy.js

aws sdk/cli

packer.iovagrant

jenkins

jenkins

code your infrastructure

code your provisioning

Summary

1. Automate as much as possible2. Find Key Metrics and monitor3. Solve problems before they happen