Stop 03 · Linux & Scripting BasicsTopics · 5/9

Python for Cloud Automation

In short

Clicking through the console to stop 30 test servers every evening gets old by day two. A 15-line Python script can do it in seconds, every day, without forgetting one.

A general-purpose language widely used for cloud automation beyond what Bash comfortably handles — most cloud providers offer a mature Python SDK (boto3 for AWS) for programmatic infrastructure control.

The blob character relaxes with tea next to a washing machine running on its own.

Think of it like this

Washing machine, not bucket

Hand-washing clothes every day works, but it's tiring and you'll skip a day. A washing machine program does the same steps the same way every time while you drink chai. Python scripts are your cloud's washing machine.

Watch it happen

Stopping dev servers with boto3

Python script

boto3

AWS EC2 API

describe_instances()

1/5Your script asks boto3 for running instances tagged env=dev.

The big ideas

What you need to know

SDKs do the heavy lifting

boto3 (AWS), google-cloud libraries (GCP) and azure-sdk (Azure) wrap every API so you call functions, not raw HTTP.

Credentials stay outside code

The SDK reads keys from your CLI profile, environment or the server's IAM role. Never paste keys into the script.

Tags make scripts safe

Filter by tags like env=dev so your script never touches production by accident.

Schedule it

Run it with cron, a CI job or as a Lambda on a timer, and the task happens even when you forget.

See it in code

Walk through the code

python

1import boto32 3ec2 = boto3.client("ec2", region_name="ap-south-1")4 5resp = ec2.describe_instances(6    Filters=[7        {"Name": "tag:env", "Values": ["dev"]},8        {"Name": "instance-state-name", "Values": ["running"]},9    ]10)11 12ids = [i["InstanceId"] for r in resp["Reservations"] for i in r["Instances"]]13 14if ids:15    ec2.stop_instances(InstanceIds=ids)16print(f"Stopped: {ids}")

1/4Create an EC2 client for the Mumbai region. boto3 picks up your credentials on its own.

Install with pip install boto3. Credentials come from aws configure or the machine's IAM role, never from the code.

Save yourself the pain

Common mistakes

Oops

Hard-coding access keys in the script and pushing it to GitHub.

Do this instead

Use aws configure locally and IAM roles on servers. Leaked keys get found and abused fast.

Oops

Testing a delete or stop script on all resources first.

Do this instead

Print what you would change first (a dry run), filter by tags, and test on one throwaway resource.

Your mission

List before you act

~35 min

Quick check

Where should a Python script running on an EC2 instance get its AWS credentials?

Go deeper

1Write a Python script using your cloud provider's SDK to list or create a resource.
2Learn to read cloud SDK documentation to find the right method for a task.
3Practice handling errors from an SDK call gracefully rather than letting the script crash silently.