AWS offers more than 200 services, which makes it easy for a beginner to get lost. You only need a handful to start: identity and permissions, servers, storage, networking, databases, load balancing, monitoring and serverless functions. This guide explains each one with a hands-on exercise, then covers certifications and the questions freshers are asked in cloud interviews. Pair it with our DevOps for Beginners guide for Linux, Git, Docker, Kubernetes and CI/CD.
Before any lab: set a billing alert on day one, use an IAM user rather than the root account, stay within the free tier, and delete test resources when you finish. AWS charges for what you use, and a forgotten resource keeps charging.
Contents
- 1. A 4-week AWS starter plan
- 2. IAM: users, groups, roles and policies
- 3. EC2: launching your first server
- 4. S3: buckets, objects and permissions
- 5. VPC: subnets, route tables and security groups
- 6. RDS: managed databases
- 7. Load balancers and Auto Scaling
- 8. CloudWatch: logs, metrics and alarms
- 9. Lambda and serverless
- 10. Which AWS certification first?
- 11. AWS interview questions for freshers
1. A 4-week AWS starter plan
- Week 1, Linux and networking: almost every cloud server runs Linux. Practise the file system, editing files, users and permissions and processes, plus IP addresses, ports, DNS and public versus private networks.
- Week 2, core AWS: learn IAM, EC2, S3 and VPC well. Launch a server, connect to it, host a simple web page, then shut everything down.
- Week 3, Git, scripting and a pipeline: put your code in Git, script a setup task, and build a basic CI/CD pipeline with Jenkins or GitHub Actions that deploys your page on every push.
- Week 4, containers: package the app with Docker, run it locally and then on EC2, and learn why teams use Kubernetes.
Keep one GitHub repository with a README and a simple architecture diagram. Walking an interviewer through one small project you built is more convincing than listing many services. Gen AI assistants can help read errors and draft scripts, but understand every line before using it.
2. IAM: users, groups, roles and policies
Identity and Access Management decides who can do what in an AWS account, and most cloud security problems start with permissions that are too broad.
- User: a person or service account that signs in.
- Group: users who share the same permissions.
- Role: permissions given temporarily to a service or trusted account, such as an EC2 instance that needs to read S3.
- Policy: a document listing which actions are allowed or denied on which resources.
Follow least privilege: start with no access and add specific actions one by one, which limits the damage of mistakes or leaked credentials. Exercise: create a user who can only read objects in one bucket, then confirm they can read but not delete.
3. EC2: launching your first server
- Choose an operating system image, such as Amazon Linux or Ubuntu.
- Pick a small, free-tier instance type.
- Create or select a key pair for secure login.
- Set a security group that allows SSH only from your own IP address.
- Connect over SSH, update the system, install a small web server and open the server’s address in a browser.
When you finish, stop or terminate the instance and delete unused volumes (EBS disks) and Elastic IP addresses so they do not keep charging.
4. S3: buckets, objects and permissions
S3 stores files of any type, from website images to logs and backups. A bucket is a container with a globally unique name in one region; an object is a file plus metadata, identified by its key, such as reports/2026/november.csv. S3 has no real folders: the “folders” in the console are just prefixes in object keys.
New buckets are private, with Block Public Access switched on; keep it on unless you are deliberately hosting a public website, because many real data leaks came from buckets made public “just for testing”. Access is controlled by IAM policies on users or roles and by bucket policies on the bucket. An application on EC2 should use an IAM role, never access keys stored in code.
Storage classes trade storage cost for retrieval cost: Standard for frequently read data, Infrequent Access and Glacier for older data. Lifecycle rules move old objects to cheaper classes or delete them automatically. Exercise: create a bucket, upload files under two prefixes, enable versioning and restore an old version, add a rule deleting temp/ objects after 7 days, and try aws s3 ls and aws s3 cp from the CLI.
5. VPC: subnets, route tables and security groups
Every EC2 server and most databases live inside a VPC, your own private network, and a misconfigured VPC is the usual cause of “my server is not reachable”.
- VPC: a private network with an IP range (a CIDR block) such as
10.0.0.0/16. Plan ranges with room to grow that do not overlap networks you must connect. - Subnets: smaller ranges, each in one Availability Zone (an isolated location within a Region).
- Internet gateway: connects the VPC to the internet.
- Route tables: decide where traffic from a subnet goes.
- NAT gateway: lets private servers reach the internet for updates without being reachable from it.
A subnet is public when its route table sends 0.0.0.0/0 to the internet gateway; a private subnet has no such route. Put load balancers in public subnets and application servers and databases in private ones. Security groups are stateful firewalls on instances with allow rules only; network ACLs apply to whole subnets, are stateless and support deny rules. Let the database security group accept traffic only from the application’s security group. Exercise: build a VPC with public and private subnets across two zones, a web server in public and a database in private. Delete NAT gateways afterwards, because they are charged by the hour.
6. RDS: managed databases
RDS runs MySQL, PostgreSQL, MariaDB, SQL Server, Oracle or Aurora for you: installation, patching, automated backups with point-in-time restore, monitoring and optional standby copies. You still design tables, write queries and manage database users.
When creating one, start with a small instance class, use general-purpose SSD storage with autoscaling, keep public access off and the database in private subnets, allow its port (3306 for MySQL) only from your app’s security group, and set a backup retention period such as 7 days. Your app connects to the endpoint hostname like any database, with the password in Secrets Manager or an environment variable.
Interview favourite: Multi-AZ keeps a standby copy in another Availability Zone and fails over automatically, for availability; read replicas are extra copies serving read queries, for performance. Turn Multi-AZ off and delete test databases while practising.
7. Load balancers and Auto Scaling
One server is a single point of failure. The standard pattern is an Application Load Balancer in front of an Auto Scaling group of EC2 instances. The ALB receives all traffic on one address, spreads requests across healthy instances in several Availability Zones, runs health checks (for example /health every 30 seconds), handles HTTPS with certificates from Certificate Manager, and can route by path.
The Auto Scaling group keeps between a minimum and maximum number of instances from a launch template (AMI, instance type and startup script), replacing any that fail health checks. Target tracking, such as keeping average CPU near 50%, is the simplest scaling policy; scheduled and step scaling are alternatives. Scaling only works if servers are stateless (sessions and uploads in a database, cache or S3), start ready to serve without manual steps, and get a sensible health-check grace period.
Lab: a launch template that installs a web server, a group of 2 to 4 instances across two zones and an ALB in front. Stop one instance and watch it be replaced, then delete everything.
8. CloudWatch: logs, metrics and alarms
CloudWatch tells you something is wrong, ideally before users do. Metrics such as EC2 CPU, RDS free storage, Lambda errors and load balancer 5xx counts arrive automatically; EC2 memory and disk need the CloudWatch agent. Logs are stored in log groups and searched with Logs Insights:
fields @timestamp, @message
| filter @message like /ERROR/
| sort @timestamp desc
| limit 20Set a retention period on every log group, because logs are kept forever by default. An alarm changes state when a metric crosses a threshold for a set time, such as CPU above 80% for 10 minutes, and can notify you through SNS. A starter setup for a small app: alarms on load balancer 5xx errors, high app-server CPU, low database storage and Lambda errors, one dashboard with these metrics, and 14 to 30 days of log retention. A few meaningful alarms beat many noisy ones.
9. Lambda and serverless
With Lambda you upload a function and AWS runs it only when triggered, charging per request and per millisecond, with no server to patch.
def lambda_handler(event, context):
name = event.get("name", "learner")
return {"statusCode": 200, "body": f"Hello, {name}"}Common triggers are API Gateway (HTTP endpoints), S3 uploads (such as processing a CSV), EventBridge schedules (cron jobs without a server) and SQS or DynamoDB streams. Know the limits: 15 minutes per run, 128 MB to 10 GB of memory, slower cold starts after quiet periods, and no persistent local state. Each function runs with an IAM execution role that should have only the permissions it needs. Lambda suits event-driven tasks, small APIs and uneven traffic; long-running or very steady heavy workloads are often simpler on EC2 or containers. Exercise: a function triggered by a CSV upload to S3 that counts rows and logs the result.
10. Which AWS certification first?
If you are new to cloud, a foundational certification covering cloud concepts, core services and billing gives you the vocabulary for everything else. After a few months of hands-on practice, an associate-level certification in a path such as solutions architecture or development shows employers you can apply that knowledge. Certification is not a substitute for projects: build one lab for each topic you study and describe it in your portfolio. Exams change, so always check the official exam guide before booking.
11. AWS interview questions for freshers
- EC2 instance vs S3 bucket? EC2 is a virtual server that runs software; S3 stores files as objects.
- IAM role vs IAM user? A user is a long-term identity for a person or service; a role grants temporary permissions to whoever assumes it, such as an EC2 instance.
- Region vs Availability Zone? A Region is a geographic area; an AZ is an isolated location within it, used to separate workloads for resilience.
- Public vs private subnet? A public subnet routes to an internet gateway; a private one does not.
- Route table vs security group? A route table decides where traffic goes; a security group decides which traffic is allowed.
- What is EBS? Block storage for EC2, such as the disk for the operating system or application data.
- What is a bucket policy? A policy attached to a bucket controlling who can access it and its objects.
- What does an S3 lifecycle rule do? Automatically moves or expires objects after a set time.
- Why least privilege? It limits the impact of mistakes and compromised credentials.
- What should you document after a lab? The architecture, commands, verification steps and clean-up, confirming temporary resources were removed.
Answer with a short example from your own project: “I deployed a small web app on EC2, stored its images in S3 and gave the app access through an IAM role instead of hard-coded keys.” For more, see the AWS and DevOps questions and answers and three cloud projects for a fresher portfolio.
Learn AWS with ScholarLogic
These services are practised in live labs in the Cloud Computing, AWS and DevOps course in Bangalore at ScholarLogic, BTM Layout; see also where to start as a fresher. Book a free demo class and a counsellor will share the fee and the next batch dates. To book, call 78923 93639.

DevOps for Beginners: Linux, Git, Docker, Kubernetes, CI/CD, Terraform and Monitoring - ScholarLogic
[…] DevOps sounds like an endless list of tools, but beginners learn it in a fixed order because each tool builds on the one before: Linux, Git, containers, orchestration, automation and monitoring. This guide follows that order with commands and small exercises for each step, and ends with the questions DevOps interviews ask. For the AWS services these tools run on, see our AWS for Beginners guide. […]
IT Career Guide for Freshers: Resumes, LinkedIn, Portfolio Projects, Interviews and Gen AI - ScholarLogic
[…] guides: Excel, SQL, Power BI, statistics, machine learning, Python, Django, Core Java, Spring Boot, AWS, DevOps and digital […]