============
explore california is a travel company website that has over 100,000 visit per month
(http://explorecalifornia.org/) and want to ship new features faster, safely & more reliably ?
How can DevOps fix this ?
Production deploys 100% automated through CI & CD pipelines.
The CI pipeline into production builds the website on the fly & run unit tests before every
pull request enters review. The CD pipeline deploys the website artifacts into AWS using terraform
a separate pipeline runs daily to merge the master branches from each dev team into an integration
environment that runs integration tests and report failures to each dev teams which makes testing easier.
All envorinoments are consistent thanks to Docker / Docker compose as devellopers can now use the same
libraries & tool chains as what's available in production.
* Writing a Dockerfile to run the website locally with Docker
* Testing the app with RSpec, Capybara, Selenium to talk to a real browser & test elements from it
* Writing Terraform code to deploy the website into an AWS s3 bucket to host the website
* Writing and running integration tests
* Writing an automated CI/CD pipeline with Jenkins that will build, test, and deploy the app.
* apply DevOps principles to a sample app for a travel booking company
## Technical implmentation step by step ##
* installed rspec, capybara, selenium-webdriver using a dockerfile & used an nginx image which will host my website
* created a docker-compose manifest & configured the below services / containers:
- the website: configured the port it will serving on & start the website locally
- Jenkins to build CI pipeline & run tests on every commit for quality releases
- awcli: used to copy the website to the aws s3 bucket
- terraform: to allow us to deploy the website onto a real env. (aws)
- selenium: configured it for other containers to have access to it too
- unit-tests: make the test available in the container
- integration-tests: to run agaisnt the copy of our website on the internet
* wrote a simple unit test using rspec, capybara & selenium to show the website logo
* set up my iam user & secret keys etc. for aws authentication
* wrote the terraform code to create an s3 bucket, configured the bucket & iam policies.
* deployed the website onto the configured bucket using the aws s3 cp command & tested the website works in the cloud
* tested the website locally.
* wrote the integration tests to run agaisnt the newly created website URL on the internet this time
* wrote a CI/CD pipeline with a jenkinsfile to automate the build, test & the website deployment
* Dockerfile: manifest that describe the image that the container will use
* RSpec: ruby based testing framework, its DSL declarative & easy to read / understand.
* Capybara: something that can spin up the website & uses a web driver to create a browser and
interact with the website e.g get the html of the website, run javascripts & test
different functionalities.
* Selenium: webdriver that spin up a real web browser and test it against whatever the web browser sees.
* Terrafrom: golang based tool used to deploy & provision any sort of infra into any sort of env
[jenkins@olivier-linux-server devops-projects]$ docker-compose run --rm unit-tests
including Capybara::DSL in the global scope is not recommended!
/usr/local/bundle/gems/selenium-webdriver-3.142.7/lib/selenium/webdriver/remote/driver.rb:39: warning: Using the last argument as keyword parameters is deprecated; maybe ** should be added to the call
/usr/local/bundle/gems/selenium-webdriver-3.142.7/lib/selenium/webdriver/remote/bridge.rb:44: warning: The called method `handshake' is defined here
.
Finished in 6.44 seconds (files took 2.8 seconds to load)
1 example, 0 failures