Hi hi, wondering if anyone can point me in the right direction here… I’m using Pulumi Deployments I’ve configured with Pulumi itself: `pulumiservice.DeploymentSettings`:

def deployment_settings(project: str, stack: str, repo_dir: str):
    return pulumiservice.DeploymentSettings(
        resource_name=f"Deployment Settings: {project}/{stack}",
        executor_context=pulumiservice.DeploymentSettingsExecutorContextArgs(
            executor_image="<http://ghcr.io/my-github-org/pulumi-deployments-executor:latest|ghcr.io/my-github-org/pulumi-deployments-executor:latest>",
        ),
        github=pulumiservice.DeploymentSettingsGithubArgs(
            deploy_commits=True,
            preview_pull_requests=True,
            repository="my-github-org/infrastructure",
            paths=[f"{repo_dir}/**"],  # Limit auto-deploy to the specified directory
        ),
        operation_context=pulumiservice.DeploymentSettingsOperationContextArgs(
            oidc=pulumiservice.OperationContextOIDCArgs(
                aws=pulumiservice.AWSOIDCConfigurationArgs(
                    duration="45m0s",
                    role_arn="arn:aws:iam::MY-ACCOUNT:role/PulumiOIDC",
                    session_name="pulumi",
                ),
            ),
            options=pulumiservice.OperationContextOptionsArgs(
                skip_install_dependencies=True,
            ),
            pre_run_commands=["prerun.sh"],
        ),
        organization="mighty",
        project=project,
        stack=stack,
        source_context=pulumiservice.DeploymentSettingsSourceContextArgs(
            git=pulumiservice.DeploymentSettingsGitSourceArgs(
                branch="refs/heads/main",
                repo_dir=repo_dir,
            ),
        ),
    )

This allows me to deploy either from the UI via the button and automatically happens during PRs to my infra repo. I’m now working on something else where I want to use the Automation API to trigger a deployment of a different stack not otherwise set up for Deployments. Looking to do something like: https://github.com/pulumi/automation-api-examples/blob/main/python/remote_deployment/main.py However I’m running into trouble with GitHub authentication. Says

failed to clone repo:: failed to create workspace, unable to enlist in git repo: unable to clone repo: authentication required; retrying

I’m curious if there’s a better way to do this. Is there a way to create a deployment configuration and trigger that?

it looks like at least the go automation sdk supports this, but I don't see a similar option for the python sdk. you might still be able to get it to work by creating deployment settings on the stack w/ the executor image config set, and then using the inherit_settings option on your

create_or_select_remote_stack_git_source

call, which I believe will cause the remote deployment to inherit non-overridden deployment settings for the stack.

ahh interesting… I’m thinking to instead use the Automation API to create a stack and then use the Deployments REST API to create a deployment

yeah I think that's probably a more straightforward approach!

Do you know if there’s any way to set a config value in the stack via the Deployment REST API? Seems like you can only set ENV variables that will be available to the deployment exectuor?

o you know if there’s any way to set a config value in the stack via the Deployment REST API? Seems like you can only set ENV variables that will be available to the deployment exectuor?

You can set config in pre-run commands by doing

pulumi config set ...

Ahh, so add another command in the

preRunCommands

list?

def trigger_pulumi_deployment(
    org: str,
    project: str,
    stack: str,
    operation: Literal["preview", "update", "destroy"],
    repo_dir: str,
    branch: str,
    custom_env: Optional[dict[str, str]] = None,
):
    url = PULUMI_API_URL.format(org=org, project=project, stack=stack)
    headers = {"Authorization": f"token {PULUMI_ACCESS_TOKEN}", "Content-Type": "application/json"}
    payload = {
        "inheritSettings": False,  # Deploy fresh settings
        "operation": operation,
        "executorContext": {
            "executorImage": EXECUTOR_IMAGE,
        },
        "github": {
            "repository": REPOSITORY,
            "deployCommits": False,
            "previewPullRequests": False,
            "pullRequestTemplate": False,
            # "paths": [],  # not needed as we do not want auto-deployments
        },
        "operationContext": {
            "oidc": {
                "aws": {
                    "duration": OIDC_DURATION,
                    "roleArn": OIDC_ROLE_ARN,
                    "sessionName": OIDC_SESSION_NAME,
                },
            },
            "options": {
                "skipInstallDependencies": True,
            },
            "preRunCommands": ["prerun.sh"],
        },
        "sourceContext": {
            "git": {
                "branch": f"refs/heads/{branch}",
                "repoDir": repo_dir,
            },
        },
    }
    if custom_env:
        payload["operationContext"]["environmentVariables"] = custom_env

    response = <http://requests.post|requests.post>(url, headers=headers, json=payload)