How do I build local commands with infrastructure as code?

In this tutorial, we’ll demonstrate how to build a local command to execute tasks directly from our infrastructure code. This can be particularly useful for running scripts or commands that need to interact with your local system or CI/CD pipelines.

First, we will define a local null_resource to represent the command execution. The null_resource is a useful abstraction that allows us to attach provisioners or triggers without requiring an actual resource to manage. For this case, we’ll use the local-exec provisioner to run our command locally.

Below is the example code block:

import * as pulumi from "@pulumi/pulumi";
import * as _null from "@pulumi/null";
import * as command from "@pulumi/command";
import * as std from "@pulumi/std";

// Define a null_resource to execute a local command
const exampleCommand = new _null.Resource("example_command", {triggers: {
    always_run: std.timestampOutput({}).apply(invoke => invoke.result),
}});
const exampleCommandProvisioner0 = new command.local.Command("exampleCommandProvisioner0", {create: "echo Hello, World!"}, {
    dependsOn: [exampleCommand],
});
export const commandOutput = "Local command executed successfully!";

Key Points:

• Provider Definition: Configures the provider required for our resources.
• null_resource: Acts as a placeholder for our local command execution.
• local-exec Provisioner: Executes the specified command on the local system.
• Triggers: Optional configuration to re-run the command based on timestamp changes.
• Output: Confirms the successful execution of our command.

Summary

By defining a null_resource with a local-exec provisioner, we can run local commands from our infrastructure as code setup. This approach enables a seamless integration of local scripts and commands within our infrastructure workflows, ensuring tasks like configuration and initial setup are executed efficiently.