Skip to main content

Creating Pipelines

Pipelines let you chain multiple deployments together and execute them in a defined order with a single trigger. This guide covers how to create, configure, and run pipelines.

When to Use a Pipeline​

Use a Pipeline when your infrastructure components depend on each other and must be provisioned in sequence, when you want a single trigger to roll out multiple pieces of infrastructure, or when you have a repeatable provisioning process across environments.

Use individual Deployments when the infrastructure is independent, or when you only need to update or rerun one specific component.

For a deeper explanation of the Pipeline model, see Pipelines.

Before Creating a Pipeline​

Pipelines reference existing Deployments — they do not create new ones. Before building a pipeline, make sure all the Deployments you want to include already exist in Amnify Deploy and each has been configured with the correct variables and cloud environment.

If a Deployment is not ready, create it first following the Creating Deployments guide.

Creating a Pipeline​

  1. Navigate to Deploy in the Amnify platform

  2. Select the cloud provider

  3. Click Pipelines in the sidebar

  4. Click New Pipeline

  5. Fill in the pipeline details:

    FieldRequiredDescription
    NameYesA descriptive name (e.g., "Full Platform Rollout — Prod")
    DescriptionNoAn explanation of what this pipeline provisions

  6. Click Create Pipeline

The pipeline is created empty. The next step is to add Deployments to it.

Adding Deployments to a Pipeline​

From the Pipeline detail page:

  1. Click Add Deployment
  2. Browse or search for the Deployments you want to include
  3. Select a Deployment and click Add
  4. Repeat for each Deployment in the pipeline

The order in which Deployments are added is their initial execution order.

Setting the Execution Order​

The order of Deployments in a Pipeline determines which one runs first, second, and so on. Each Deployment must complete successfully before the next one starts.

To reorder Deployments:

  1. From the Pipeline detail page, find the Deployments list
  2. Use the up/down arrows (or drag-and-drop if enabled) to reorder them
  3. Changes take effect on the next run

Design your order based on dependencies:

  • Provision foundational infrastructure first (networks, shared services)
  • Provision dependent infrastructure next (applications, databases that need the network)
  • Provision consumer infrastructure last (anything that connects to the above)

Running a Pipeline​

To execute all Deployments in the pipeline in sequence:

  1. Open the Pipeline detail page
  2. Click Run Pipeline
  3. A new Pipeline Run is created

The pipeline executes each Deployment in order. Each Deployment still goes through its own approval workflow — you will be prompted to approve the plan and apply for each one in turn.

Monitoring a Pipeline Run​

From the Pipeline detail page, click on the active or any past Pipeline Run to open the Pipeline Run detail view. It shows the overall pipeline status, which Deployment is currently executing (highlighted), the status of each completed Deployment in the sequence, and links to each individual Deployment Run for log access and output review.

If a Deployment fails, the pipeline stops at that point and subsequent Deployments are not executed. Fix the issue with the failing Deployment and run the pipeline again — already-completed Deployments will not be re-executed if their infrastructure is already in the correct state.

Editing a Pipeline​

You can modify a pipeline at any time:

  • Rename or update the description — click Edit on the pipeline detail page
  • Add new Deployments — click Add Deployment
  • Remove Deployments — use the remove option next to any Deployment in the list
  • Reorder Deployments — use the up/down controls

Changes to the pipeline configuration only affect future runs. Historical runs are unaffected.

Removing a Deployment from a Pipeline​

Removing a Deployment from a Pipeline does not delete the Deployment itself — it only removes it from the pipeline's execution sequence. The Deployment continues to exist and can still be executed individually or added to other pipelines.

Next Steps​