Skip to main content

Create environments

Environments represent your deployment targets (QA, Prod, etc). Each environment contains one or more Infrastructure Definitions that list your target clusters, hosts, namespaces, etc.

Create an environment​

You can create environments from:

  • Within a pipeline
  • Outside a pipeline
  • An account
  • An Organization

To create an environment from inside of a pipeline, select New Environment in the Infrastructure tab of a new CD stage.

Define the environment configuration​

In the environment Configuration, you can manage the Name, Description, Tags, and Environment Type of the environment.

You can also set default manifests, specifications, config files, and variables to use whenever Harness deploys a service to this environment.

For example, a stage has a Kubernetes service with a manifest but whenever that service is deployed to the QA environment, the manifest in that environment's Configuration overwrites the namespace of with the manifest in the service with QA.

Create service overrides​

Service overrides are different from Environment Configuration in the following ways:

  • Environment Configuration: applies to every service that is used with the environment.
  • Environment Service Overrides: applies to specific services you select. Whenever that service is used with that environment, the Service Override is applied.

Override priority​

When you are using environment configuration and service override to override service settings, it's important to understand the priority of the overrides.

The priority from top to bottom is:

  1. Environment service overrides
  2. Environment configuration
  3. Service settings

Overriding values.yaml​

You can specify values YAML files at the environment's Service Overrides and Configuration, and the service itself.

Here is an example of specifying it at the environment's Configuration:

When you have a values yaml file at two or more of the environment Service Overrides, Environment Configuration, and the service itself, Harness merges the files into a single values YAML for deployment. This merging is performed at pipeline execution runtime.

Overriding occurs when the higher priority setting has the same name:value pair as a lower priority setting.

Let's look at two examples.

Merging values.yaml name:value pairs​

An environment's Service Overrides values YAML has the name:value pair servicePort: 80 but no replicas name:value.

A service's Service Definition has a values YAML with replicas: 2 but no servicePort name:value.

At runtime, the two values YAML files are merged into one.

The servicePort: 80 from the environment Service Overrides values YAML is merged with the Service Definition's replicas: 2 in the values YAML:

Fully overriding values.yaml name:value pairs​

An environment's Service Overrides values YAML has the name:value pairs replicas: 2 and servicePort: 80. 

A service's Service Definition has a values YAML with replicas: 4 and servicePort: 8080. 

At runtime, the name:value pairs from the environment Service Overrides values YAML fully override the service values YAML. The replicas: 2 and servicePort: 80 from the environment Service Overrides are used.

Fully overriding config files and variables​

Config files are a black box that can contain multiple formats and content, such as YAML, JSON, plain text, etc. Consequently, they cannot be overridden like Values YAML files.

Variables cannot be partially overridden either. They are completely replaced.

When you have Config files at two or more of the environment Service Overrides, Configuration, and the service itself, the standard override priority is applied.

When you have Variables with the same name at two or more of the environment Service Overrides, Configuration, and the service itself, the standard override priority is applied.

Add infrastructure definitions​

Infrastructure definitions represent an environment's infrastructures physically. They are the actual clusters, hosts, namespaces, etc, where you are deploying a service.

An environment can have multiple Infrastructure Definitions. 

When you select an environment in a stage, you can select the Infrastructure Definition to use for that stage.

Propagating environments through multiple stages​

note

Currently, this feature is behind the feature flag CDS_ENV_PROPAGATION. Contact Harness Support to enable the feature.

When modeling multiple Deploy stages in a pipeline, you can propagate the environment and infrastructure definition selected in one stage to one or more subsequent stages.

When you propagate an environment, you can either use the same infrastructure definition that was used in the parent stage or you can select a different infrastructure definition.

note

You can also propagate services between stages. For more information, go to Propagate CD services.

Important notes​

  • Propagation is only supported for Deploy stages. Custom stages do not have environments.

  • You cannot propagate environments between different deployment types. For example, you cannot propagate a Kubernetes environment between a Kubernetes deployment stage and a Shell Script deployment stage.

  • Environment propagation is not supported when using multiple environments in a single stage (multi environment deployments).

  • Environment propagation is progressive. You can only propagate environments from stage to stage in a forward direction in your pipeline. For example, Stage 2 cannot propagate an environment from a subsequent Stage 3.

  • In a pipeline's Advanced Options, in Stage Execution Settings, you can set up selective stage executions. This allows you to select which stages to deploy at runtime.

    • If you select a stage that uses a propagated environment (a child environment), that stage will not work. This is because the parent environment's settings must be resolved as part of the deployment.

  • When propagation is set up between a parent stage and child stage, moving the parent or child stage out of sequence resets any propagated settings to their defaults. If you do this, you are prompted to confirm. If you confirm, the stages are reset to their defaults.

  • You cannot propagate environment from a stage which also propagates environment from another stage.

    picture 0

Propagate an environment​

  1. Open a pipeline that contains at least one Deploy stage.

  2. Add a subsequent Deploy stage.

  3. In Service, select a service for the stage, and then select Continue.

  4. In Environment, select Propagate Environment From.

  5. In Propagate Environment From, select the environment of a previous stage.

    The environment and infrastructure definition from the previous stage is now configured in this stage.

Select a different infrastructure when propagating environment from a previous stage​

When you propagate an environment from a previous stage, you have the option to select a different infrastructure definition.

note

Currently, this feature is behind the feature flag CDS_SUPPORT_DIFFERENT_INFRA_DURING_ENV_PROPAGATION. Contact Harness Support to enable the feature.

  1. Select a stage for which you want to propagate the environment and infrastructure. Make sure that the selected stage has at least one previous Deploy stage.

  2. In the Environment tab, select Propagate Environment From and select the environment of a previous stage.

  3. Select Deploy to Different Infrastructure and select an infrastructure. This option allows you to select a different infrastructure definition.

  4. Select + New Infrastructure to create a new infrastructure definition for use in your stage.

Limitation
  • Harness doesn't support nested propagation. For example, if Stage 2 is propagated from Stage 1, you cannot propagate Stage 3 from Stage 2.
  • This feature is not supported when deploying to multiple environments or infrastructures.

Here's a sample YAML snipped when a different infrastructure definition is propagated:

environment:
  useFromStage:
    stage: s1
  infrastructureDefinitions:
    - identifier: infra_1
      inputs:
        identifier: infra_1
        type: KubernetesDirect
        spec:
          connectorRef: <+input>

Define GitOps clusters​

When you use Harness GitOps you can add GitOps clusters to an environment. 

To learn more about Harness GitOps, go to Harness GitOps basics.

Next, when you create a pipeline, you can select the environment and the GitOps cluster(s) to use.

GitOps clusters are used in a PR pipeline. A PR pipeline creates and merges a Git PR on the config.json for a destination cluster as part of an ApplicationSet. The PR Pipeline runs, merges a change to the config.json, and a GitOps sync on the ApplicationSet is initiated.

GitOps Clusters are not used in standard CD pipelines. They're used when using GitOps only.

Runtime inputs and expressions in environments​

If you use runtime inputs in your environments, you will need to provide values for these when they run pipeline using these environments.

If you use expressions in your environments, Harness must be able to resolve these expressions when users run pipeline using these environments.

Select Runtime input for the environment.

When you run the pipeline, you can select the environment for their runtime inputs.

For more information on runtime inputs and expressions, go to fixed values, runtime inputs, and expressions.