Operate Bruin Cloud From the Terminal

What

You already run pipelines in Bruin Cloud. This guide moves the day-to-day operations into your terminal with the bruin cloud command, so you can check status, diagnose failures, trigger backfills, and talk to your agents without switching to the dashboard.

By the end you will be able to:

• Authenticate the CLI with a team API token, then list teams and projects
• Inspect pipelines, runs, and asset instances, and read them as plain text or json
• Diagnose a failed run in one command, read the failing logs, and rerun only what failed
• Trigger runs and split a date range into a backfill of batched runs
• Chat with a Bruin Cloud AI agent and explore warehouse cost, all from the shell

Everything here talks to the same Bruin Cloud you already use. The CLI is a second door into it, not a separate system - a run you trigger from the terminal shows up on the pipeline page, and a token you revoke in the UI stops working immediately.

How

The bruin cloud command groups its subcommands by object: teams, projects, pipelines, runs, instances, backfills, agents, and cost. Almost every command takes a --project-id, so the first step is getting authenticated and finding your IDs. After that each step follows a real operating loop: look at what ran, figure out why it broke, run it again, and ask an agent when you want a second opinion.

Two things are worth knowing up front. Add --output json (or -o json) to any command to get machine-readable output for scripts and CI. And the CLI resolves your API token in a fixed order - the --api-key flag first, then the BRUIN_CLOUD_API_KEY environment variable, then a bruin connection in .bruin.yml - so you can keep credentials out of your shell history once you settle on a setup.