In this post I’ll show how to setup Azure VMs turn off & on by schedule using Azure Automation which is really useful for cost saving.
We’ll use dozen of actions to manage this task.
However, not all actions can be managed with PowerShell, so we’ll use old Azure portal a few times.
Initially I used one big script, but for this demo I decided to split it to make things simple.
In this particular example I’ll use two VMs running in the same cloud service.
To restore file from one of the previous commits or tags do this:
We’ve completed infrastructure setup using “infrastructure as code” principles.
As we used real cloud power we’ve achieved awesome availability, scaling and security.
We’ve saved a lot of time because we haven’t setup linux, nginx, zabbix, chef, etc.
Instead of this we were focused on application, testing, scaling, continuous delivery – stuff that makes business healthy and happy.
We’ve saved a lot of pretty expensive DevOps hours and haven’t spent too much money because AWS EB, ELB, Scaling, etc are free to use!
We’ll pay only for EC2, EBS and CloudWatch.
And because of scaling we’ll pay only for what we’ll use.
According to the requirements, we need to build and deploy a new code automatically for dev & stg.
But Prod will be deployed manually after tests performing on dev & stg.
There’re some general ideas in all scripts, so let’s overview ’em:
From build machine we’ll use AWS, BitBucket and DockerHub and I want to automate auth process.
Now it’s time to setup build machine.
I’ll use Ubuntu instead of Amazon Linux because of problems with AWS CLI installing.
Hope it will be fixed soon, but Ubuntu is quiet good for us too.
It might be a bit different from “classic” definition, but let’s call build & deploy logical steps “pipelines”.
Every logical step is put in separate function to make things simple & clear.
I’ll list and describe logical steps (functions sequence) for all actions for you:
In this part I’ll create quite simple app and this is “developer” piece of work.
We need to install node, npm & meteor on work machine:
brew install nodejs curl https://install.meteor.com/ | sh
