Getting started with CI
Shippable's DevOps Automation platform gives you an easy way to set up Continuous Integration (CI) for your projects and automate unit testing, packaging, and deployment for any change in your source control repository.
This page explains what CI is, and also the basic concepts you should know before starting to use Shippable.
Why do you need continuous integration?
Continuous Integration is an essential first part of a continuous delivery workflow and helps software teams ensure that their changes are built and tested with the latest version of the entire codebase. As a result, most bugs are found almost immediately after the code change is committed, leading to better quality since each bug can be easily isolated to a specific code change and fixed promptly.
For more on Continuous Integration and why you should include it as part of your workflow, read Martin Fowler's article on the Benefits of Continuous Integration
The picture below shows a very basic CI workflow. Shippable receives an incoming webhook from your source control and spins up a build machine. A Shippable agent comes up on the build machine and starts the build container, inside which your CI commands are executed. At the end of the build process, you can push to any endpoint. This endpoint can be an artifact repository like Artifactory or Docker Hub, or can be a PaaS/IaaS/Container Service endpoint.
Your CI workflow stops after deployment to an endpoint. If you want to define your end to end application delivery pipelines, check out our Continuous Delivery tutorials which show you how to configure multi-stage deployments, release mgmt with semantic versioning and approval gates, and our IT Ops section which shows you how to make provisioning and infrastructure management activities part of your workflows.
Your CI workflow is configured with a YAML based file called shippable.yml which should be committed to the root of the repository you want to build. This is mandatory for all enabled projects and tells us what the build should do. For the shippable.yml structure and how to configure it, check out our Build Configuration page.
When a repository is enabled on Shippable, we enable webhooks on that repository and start listening to commit and pull request events. To enable a repository, read the Enable your project page.
Shippable automatically builds and tests your repositories when the following triggers are received-
- commit webhook
- webhook for a pull request opened for an enabled repository
- webhook for a git tag push event (Turned off by default)
- webhook for a release event (Turned off by default)
- a new version of an
INjob or resource is detected (only if you've made your CI job part of an Assembly Line workflow)
You can also initiate manual builds through the UI, by clicking on the Build button for any project or branch, irrespective of a webhook event.
To learn how to switch some of these triggers off, read about the configuration in Triggering your CI.
When a build is triggered, it is executed in the sequence below -
- First, we provision a build machine for you. Build machines have 2 cores, 3.75 GB RAM each. Provisioning can take anytime between 10 seconds to 4 mins, depending on whether your subscription already has a build machine running from an earlier build.
- Once the machine is available, the Shippable agent starts running on that machine.
- Commands in the
pre_cisection are then executed on your build machine by the Shippable Agent.
- The next step is booting your build container. This will use our default Docker images if nothing is configured in the
pre_ci_bootsection of your yml. If that section is configured, it overrides the default image and boots up the build container specified.
- Next, we set the environment in the build container and clone your repository that is to be built.
- All commands in the
cisection are executed in sequence inside the build container.
- Commands in the
post_cisection are executed inside the build container.
- If the
post_cisections were successful, commands in the
on_successsection are executed.
- If the
post_cisections failed, commands in the
on_failuresection are execured.
- If notifications are configured in the yml, we will send out notifications about build results through the configured channel. Email notifications are on by default.
Configuring Continuous Delivery
Continuous integration is a very developer-focused workflow, but you also need to automate a bunch of other activities to achieve Continuous Delivery, such as deployment into multiple environments (e.g. Test, Staging, Production), as well as running test suites, provisioning infrastructure, etc.
If you want to use Shippable to configure these end to end workflows, the following documents will be useful: