Auto DevOps

Introduced in GitLab 11.0.

GitLab Auto DevOps is a collection of pre-configured features and integrations that work together to support your software delivery process.

Auto DevOps features and integrations:

  • Detect your code’s language.
  • Build and test your application.
  • Measure code quality.
  • Scan for vulnerabilities and security flaws.
  • Check for licensing issues.
  • Monitor in real time.
  • Deploy your application.

For an introduction to Auto DevOps, watch Auto DevOps in GitLab 11.0.

Auto DevOps features

Based on the DevOps stages, use Auto DevOps to:

Build your app:

Test your app:

Deploy your app:

Monitor your app:

Secure your app:

How it works

Auto DevOps detects your code language and uses CI/CD templates to create and run default pipelines. All you need to kick it off is to enable it.

Auto DevOps starts by building and testing your application. Then, based on your predefined deployment configuration, creates the necessary jobs to deploy your apps to staging and/or production. It also sets up Review Apps so that you can preview your changes in a per-branch basis.

Note that you don’t need to set up the deployment upfront. Auto DevOps still builds and tests your application. You can define the deployment later.

Auto DevOps avoids the hassle of having to create entire pipelines manually. Keep it simple and facilitate an iterative approach: ship your app first, then explore the customizations later. You can also manage Auto DevOps with APIs.

Some of the benefits of using Auto DevOps as part of your workflow are:

  • Consistency: always start from default templates.
  • Simplicity: create your pipeline with the default settings first, iterate later.
  • Productivity: deploy multiple apps in a short period of time.
  • Efficiency: get things done fast.

Comparison to application platforms and PaaS

Auto DevOps provides features often included in an application platform or in a Platform as a Service (PaaS).

Inspired by Heroku, Auto DevOps goes beyond it in multiple ways:

  • Auto DevOps works with any Kubernetes cluster.
  • There is no additional cost.
  • You can use a cluster hosted by yourself or on any public cloud.
  • Auto DevOps offers an incremental graduation path. If you need to customize, start by changing the templates and evolve from there.

Get started with Auto DevOps

To get started, you only need to enable Auto DevOps. This is enough to run an Auto DevOps pipeline to build and test your application.

If you want to build, test, and deploy your app:

  1. See the requirements for deployment.
  2. Enable Auto DevOps.
  3. Follow the quick start guide.

As Auto DevOps relies on many components, be familiar with:

When deploying to a Kubernetes cluster make sure you’re also familiar with:

Enable or disable Auto DevOps

Version history

Depending on your instance type, you can enable or disable Auto DevOps at the following levels:

Instance type Project Group Instance (Admin Area)
GitLab SaaS Yes No No
GitLab self-managed Yes Yes Yes

Before enabling Auto DevOps, consider preparing it for deployment. If you don’t, Auto DevOps can build and test your app, but cannot deploy it.

At the project level

To use Auto DevOps for individual projects, you can enable it in a project-by-project basis. If you intend to use it for more projects, you can enable it for a group or an instance. This can save you the time of enabling it one by one.

Only project Maintainers can enable or disable Auto DevOps at the project level.

Before enabling Auto DevOps, ensure that your project does not have a .gitlab-ci.yml present. If present, your CI/CD configuration takes precedence over the Auto DevOps pipeline.

To enable Auto DevOps for a project:

  1. Go to your project’s Settings > CI/CD > Auto DevOps.
  2. Select the Default to Auto DevOps pipeline.
  3. (Recommended) Add the base domain.
  4. (Recommended) Choose the deployment strategy.
  5. Select Save changes.

GitLab triggers the Auto DevOps pipeline on the default branch.

To disable it, follow the same process and deselect Default to Auto DevOps pipeline.

At the group level

Introduced in GitLab 11.10.

When you enable Auto DevOps at group level, the subgroups and projects in that group inherit the configuration. This saves you time by batch-enabling it rather than enabling individually for each subgroup or project.

When enabled for a group, you can still disable Auto DevOps for the subgroups and projects where you don’t want to use it.

Only GitLab administrators and group owners can enable or disable Auto DevOps at the group level.

To enable Auto DevOps for a group:

  1. Go to your group’s Settings > CI/CD > Auto DevOps.
  2. Select Default to Auto DevOps pipeline.
  3. Select Save changes.

After enabling Auto DevOps at the group level, you can trigger the Auto DevOps pipeline for any project that belongs to that group. To do so:

  1. Go to the project’s homepage.
  2. Make sure the project doesn’t contain a .gitlab-ci.yml file.
  3. From the project’s sidebar, go to CI/CD > Pipelines.
  4. Select Run pipeline to trigger the Auto DevOps pipeline.

To disable Auto DevOps on the group level, follow the same process and deselect Default to Auto DevOps pipeline.

At the instance level

By enabling Auto DevOps in the instance level, all projects created in that instance become enabled. This is convenient when you want to run Auto DevOps by default for all projects. You can still disable Auto DevOps individually for the groups and projects where you don’t want to run it.

Only GitLab administrators can enable or disable Auto DevOps in the instance level.

Even when disabled for an instance, group owners and project maintainers can still enable Auto DevOps at the group and project levels.

To enable Auto DevOps for your instance:

  1. From the top bar, select Menu > Admin.
  2. Go to Settings > CI/CD > Continuous Integration and Deployment.
  3. Select Default to Auto DevOps pipeline.
  4. (Optional) Add the Auto DevOps base domain.
  5. Select Save changes.

When enabled, it attempts to run Auto DevOps pipelines in every project. If the pipeline fails in a particular project, it disables itself. GitLab administrators can change this in the Auto DevOps settings.

If a CI/CD configuration file is present, it remains unchanged and Auto DevOps doesn’t affect it.

To disable Auto DevOps in the instance level, follow the same process and deselect the Default to Auto DevOps pipeline checkbox.

Quick start

To guide your through the process of setting up Auto DevOps to deploy to a Kubernetes cluster on Google Kubernetes Engine (GKE), see the quick start guide.

You can also follow the quick start for the general steps, but deploy to AWS ECS instead.

If you’re a self-managed user, before deploying to GKE, a GitLab administrator needs to:

  1. Configure the Google OAuth 2.0 OmniAuth Provider.
  2. Configure a cluster on GKE.

Upgrade Auto DevOps dependencies when updating GitLab

When updating GitLab, you may need to upgrade Auto DevOps dependencies to match your new GitLab version:

Limitations

Private registry support

We cannot guarantee that you can use a private container registry with Auto DevOps.

We strongly advise you to use GitLab Container Registry with Auto DevOps to simplify configuration and prevent any unforeseen issues.

Install applications behind a proxy

The GitLab integration with Helm does not support installing applications when behind a proxy.

To do so, inject proxy settings into the installation pods at runtime. For example, you can use a PodPreset:

apiVersion: settings.k8s.io/v1alpha1
kind: PodPreset
metadata:
  name: gitlab-managed-apps-default-proxy
  namespace: gitlab-managed-apps
spec:
  env:
    - name: http_proxy
      value: "PUT_YOUR_HTTP_PROXY_HERE"
    - name: https_proxy
      value: "PUT_YOUR_HTTPS_PROXY_HERE"

Troubleshooting

See troubleshooting Auto DevOps.