Why This Matters

DevOps engineers live in terminals. The CLI brings Copilot to where they work — without forcing them into an editor. This unlocks AI assistance for ops, infra, and production debugging.

Concept Explained

The CLI is ideal for terminal-native workflows: debugging production issues, writing infra scripts, executing ops runbooks, analyzing logs. No editor needed.

How It Works

SSH into a production host. Run copilot. Ask about logs, errors, or commands. Copilot reads terminal output, proposes fixes or commands. Apply selectively.

Step-by-Step Tutorial

1. Identify terminal workflow

2. Run copilot in terminal

3. Use @terminal

4. Apply selectively

5. Document runbooks

Real-World Example

A DevOps engineer SSH'd into a production host with a failing service. Ran copilot. Asked: 'the service log shows OOM errors, what should I check?'. Copilot listed: memory limits, recent deployments, memory leaks. Engineer checked each, found a memory leak introduced in the last deploy. Rolled back. Five-minute resolution.

Example Prompts / Commands / Code

# Production debugging:
$ ssh prod-host
$ copilot
> @terminal the auth service is returning 500s, check the logs
[Copilot reads recent log output, proposes diagnosis]

# Infra script:
$ copilot
> write a script to find all EC2 instances older than 30 days
[Copilot writes AWS CLI script]

# Ops runbook:
$ copilot
> we have a database connection leak, what's the runbook?
[Copilot proposes steps: check connections, restart pool, alert team]

# Log analysis:
$ copilot
> analyze the last 1000 lines of /var/log/app.log, find error patterns
[Copilot identifies common errors, suggests causes]

Before running Copilot-suggested commands in production:
[ ] Command is read-only (ls, cat, grep, ps)?
[ ] If write command, reviewed by a human?
[ ] Tested in staging first?
[ ] Documented in runbook?
[ ] Rollback plan ready?

Never let Copilot auto-apply destructive commands in production.

Common Mistakes

• Letting Copilot auto-apply destructive commands in production.
• Not testing in staging first.
• Forgetting production safety rules.
• Using CLI when an editor would be safer (for complex code changes).

Best Practices

• Use CLI for terminal-native ops workflows.
• Always review proposed commands before running in production.
• Test in staging first when possible.
• Document successful workflows as runbooks.
• Never auto-apply destructive commands in production.

Troubleshooting

Practical Exercise

Professional Challenge

Frequently Asked Questions

Further Reading

• CLI reference

Official References

• CLI reference

Related lessons: BE-42, IN-41