Cloud Free Tier Guide: Maximize Free Resources for Certification Study¶

A comprehensive guide to leveraging free tier offerings from major cloud providers (AWS, Azure, GCP, Oracle, IBM) for certification preparation and hands-on practice.

Table of Contents¶

• AWS Free Tier
• Microsoft Azure Free Tier
• Google Cloud Platform Free Tier
• Oracle Cloud Free Tier
• IBM Cloud Free Tier
• Cost Management Strategies
• Free Tier Best Practices
• Multi-Cloud Strategy
• Study Projects Using Free Tiers
• Common Pitfalls and How to Avoid Them

AWS Free Tier¶

Website: https://aws.amazon.com/free/

Duration: 12 months from account creation + Always Free services

Credit Card Required: Yes

Free Tier Categories¶

1. 12 Months Free (from account creation) 2. Always Free (no expiration) 3. Free Trials (short-term trials for specific services)

12 Months Free Services¶

Compute¶

EC2 (Elastic Compute Cloud) - Limit: 750 hours per month of t2.micro or t3.micro instances (Linux, RHEL, or SLES) - What You Can Do: - Run one t2.micro instance 24/7 for a full month - Or run multiple instances that total 750 hours - Practice: Deploy web servers, application servers, practice SSH access - Watch Out: Data transfer limits apply (15GB out per month)

Elastic Load Balancing - Limit: 750 hours per month - Limit: 15GB data processing - What You Can Do: - Set up Application or Network Load Balancer - Practice: Load balance between multiple EC2 instances

Storage¶

S3 (Simple Storage Service) - Limit: 5GB of standard storage - Limit: 20,000 GET requests - Limit: 2,000 PUT requests - What You Can Do: - Host static websites - Store files, images, backups - Practice: Bucket policies, lifecycle rules, versioning

EBS (Elastic Block Store) - Limit: 30GB of General Purpose (SSD) or Magnetic storage - Limit: 2 million I/Os (with Magnetic) - Limit: 1GB of snapshot storage - What You Can Do: - Attach volumes to EC2 instances - Practice: Volume management, snapshots, encryption

Elastic File System (EFS) - Limit: 5GB of standard storage - What You Can Do: - Shared file system for EC2 - Practice: Mount EFS on multiple instances

Database¶

RDS (Relational Database Service) - Limit: 750 hours per month of db.t2.micro, db.t3.micro, or db.t4g.micro instances - Limit: 20GB of General Purpose (SSD) storage - Limit: 20GB of backup storage - Supported: MySQL, PostgreSQL, MariaDB, Oracle BYOL, SQL Server Express - What You Can Do: - Deploy MySQL, PostgreSQL databases - Practice: Backups, snapshots, read replicas, parameter groups

DynamoDB - Limit: 25GB of storage - Limit: 25 read/write capacity units - What You Can Do: - Build NoSQL applications - Practice: Table design, queries, indexes, streams

ElastiCache - Limit: 750 hours per month of cache.t2micro or cache.t3.micro nodes

Networking¶

VPC (Virtual Private Cloud) - Free: No charge for VPC creation - Free: 1 million VPN connection hours (first 12 months) - What You Can Do: - Create custom networks - Practice: Subnets, route tables, security groups, NACLs

CloudFront (CDN) - Limit: 1TB data transfer out - Limit: 10,000,000 HTTP/HTTPS requests - What You Can Do: - Set up CDN for websites - Practice: Distributions, cache behaviors, origins

Route 53 (DNS) - Limit: 50 queries per month (Hosted Zone) - Note: Hosted zone costs $0.50/month after free queries

Serverless¶

Lambda - Limit: 1 million free requests per month - Limit: 400,000 GB-seconds of compute time per month - What You Can Do: - Build serverless applications - Practice: Functions, triggers, event sources, layers - Example: Lambda function running 128MB for 1 second = 0.125 GB-seconds

API Gateway - Limit: 1 million API calls per month (first 12 months)

Other Services¶

SNS (Simple Notification Service) - Limit: 1,000 publishes - Limit: 100,000 HTTP/S deliveries - Limit: 1,000 email deliveries

SQS (Simple Queue Service) - Limit: 1 million requests per month

CloudWatch - Limit: 10 custom metrics and 10 alarms - Limit: 1 million API requests - Limit: 5GB of log data ingestion and storage - Limit: 3 dashboards with up to 50 metrics each

AWS CodeBuild - Limit: 100 build minutes per month

AWS CodePipeline - Limit: 1 active pipeline per month

Always Free Services¶

Lambda: 1 million requests + 400,000 GB-seconds per month (forever) DynamoDB: 25GB storage + 25 read/write capacity units (forever) SNS: 1 million publishes (forever) SQS: 1 million requests (forever) CloudWatch: 10 metrics, 10 alarms, 1 million API requests (forever) Cognito: 50,000 monthly active users AWS Glue: 1 million objects stored in the Data Catalog AWS Step Functions: 4,000 state transitions per month

Free Trials¶

Amazon SageMaker: 250 hours per month of t2.medium notebook usage (2 months) Amazon Redshift: 750 hours per month of dc2.large node usage (2 months) Amazon Inspector: 90-day trial Amazon GuardDuty: 30-day trial AWS Config: Free during first 30 days

Data Transfer Limits (Critical)¶

Free Inbound Data Transfer: Unlimited to AWS Free Outbound Data Transfer: 15GB per month (aggregated across all services) Exceeding Limits: $0.09/GB for the next 10TB

Important: Data transfer is one of the most common unexpected costs.

Microsoft Azure Free Tier¶

Website: https://azure.microsoft.com/free/

Duration: 12 months + $200 credit (30 days) + Always Free services

Credit Card Required: Yes

Phone Verification Required: Yes

$200 Credit (First 30 Days)¶

• Use for any Azure service
• Expires after 30 days or when credit exhausted
• Good for experimentation with expensive services
• Best use: Test services not in free tier (AKS clusters, larger VMs)

12 Months Free Services¶

Compute¶

Virtual Machines (Linux) - Limit: 750 hours per month of B1S (1 vCPU, 1GB RAM) - What You Can Do: - Run one B1S VM continuously - Practice: VM deployment, extensions, availability sets

Virtual Machines (Windows) - Limit: 750 hours per month of B1S - What You Can Do: - Run Windows Server - Practice: Active Directory, IIS, PowerShell

Container Instances - Limit: 50 container group deployments - What You Can Do: - Deploy containerized applications - Practice: Container management without Kubernetes

Storage¶

Blob Storage - Limit: 5GB locally redundant storage (LRS) - Limit: 20,000 read operations - Limit: 10,000 write operations - What You Can Do: - Store files, backups, static websites - Practice: Blob tiers, lifecycle management

File Storage - Limit: 100GB LRS - What You Can Do: - SMB file shares - Practice: File share mounting on VMs

Managed Disks - Limit: 2 x 64GB P6 SSD disks - Limit: 2 x 128GB P10 SSD disks - What You Can Do: - Attach to VMs - Practice: Disk management, snapshots

Database¶

SQL Database - Limit: 250GB - What You Can Do: - Deploy SQL Server databases - Practice: Backups, geo-replication, elastic pools

Cosmos DB - Limit: 25GB storage - Limit: 400 request units per second - What You Can Do: - Build multi-model NoSQL applications - Practice: Global distribution, consistency levels

Networking¶

Load Balancer - Limit: 750 hours - What You Can Do: - Load balance VMs - Practice: Health probes, load balancing rules

Virtual Network - Limit: 50 virtual networks - Free: No charge for VNets - What You Can Do: - Create private networks - Practice: NSGs, peering, VPN gateways

Public IP Address - Limit: 750 hours of static IPv4 - What You Can Do: - Assign to VMs, load balancers - Practice: IP management

Bandwidth - Limit: 100GB outbound data transfer - What You Can Do: - More generous than AWS (15GB)

Application Services¶

App Service - Limit: 10 web, mobile, or API apps - Limit: 1GB storage - What You Can Do: - Deploy web applications - Practice: Deployment slots, scaling

Functions - Limit: 1 million requests per month - Limit: 400,000 GB-s execution time - What You Can Do: - Build serverless applications - Practice: Triggers, bindings, Durable Functions

Azure DevOps - Limit: 5 users with Basic access - Limit: Unlimited private Git repos - Limit: 1,800 minutes of CI/CD per month - What You Can Do: - Set up CI/CD pipelines - Practice: Azure Pipelines, Repos, Boards

Always Free Services¶

App Service: 10 apps with 1GB storage Functions: 1 million executions per month Azure Cosmos DB: 1000 request units/sec with 25GB storage (first 400 RU/s free) Service Bus: 750 hours per month Event Grid: 100,000 operations per month Azure DevOps: 5 users Azure Active Directory: 50,000 stored objects Notification Hubs: 1 million pushes Azure Container Registry: 1 registry with 10GB storage Azure Kubernetes Service (AKS): Free cluster management (pay only for VMs)

Azure for Students¶

No Credit Card Required

$100 Credit (12 months) for students Free Services: Many free services without credit card Verification: Requires .edu email or other student verification

Google Cloud Platform Free Tier¶

Website: https://cloud.google.com/free

Duration: 90 days $300 credit + Always Free tier

Credit Card Required: Yes (not charged during trial)

$300 Credit (90 Days)¶

• Use for any GCP service
• Expires after 90 days or when credit exhausted
• Won't auto-charge after expiration
• Best for: Testing expensive services (GKE, larger instances)

Always Free Tier¶

No Time Limit - These services are always free within limits

Compute¶

Compute Engine - Limit: 1 e2-micro instance per month (US regions only: us-west1, us-central1, us-east1) - Specs: 0.25 vCPU, 1GB RAM - Limit: 30GB-months standard persistent disk - Limit: 1GB network egress per month (excluding China and Australia) - What You Can Do: - Run small VMs 24/7 - Practice: Instance management, metadata, startup scripts

Cloud Functions - Limit: 2 million invocations per month - Limit: 400,000 GB-seconds, 200,000 GHz-seconds compute time - Limit: 5GB network egress per month - What You Can Do: - Build serverless functions - Practice: HTTP functions, event-driven functions

Cloud Run - Limit: 2 million requests per month - Limit: 360,000 GB-seconds memory - Limit: 180,000 vCPU-seconds - Limit: 1GB network egress per month - What You Can Do: - Deploy containerized apps - Practice: Serverless containers

App Engine - Limit: 28 frontend instance hours per day - Limit: 9 backend instance hours per day - Limit: 1GB egress per day - What You Can Do: - Deploy web applications - Practice: App deployment, versions, traffic splitting

Google Kubernetes Engine (GKE) - Free: 1 Autopilot or Zonal cluster per month (cluster management fee waived) - Note: Still pay for worker nodes

Storage¶

Cloud Storage - Limit: 5GB per month (regional storage, US only) - Limit: 5,000 Class A operations per month - Limit: 50,000 Class B operations per month - Limit: 100GB network egress per month - What You Can Do: - Store files, host static websites - Practice: Buckets, lifecycle management, IAM

Cloud Firestore - Limit: 1GB storage - Limit: 50,000 reads, 20,000 writes, 20,000 deletes per day - What You Can Do: - Build NoSQL applications - Practice: Collections, queries, real-time updates

Cloud Datastore (legacy, being replaced by Firestore) - Limit: 1GB storage - Limit: 50,000 reads, 20,000 writes, 20,000 deletes per day

Networking¶

VPC (Virtual Private Cloud) - Free: VPC creation, subnets, firewall rules - Limit: Egress as noted in other services

Cloud Load Balancing - Pricing: Not in always free tier - During trial: Use $300 credit

Cloud NAT - Pricing: Not in always free tier

Databases¶

Cloud SQL - Pricing: Not in always free tier - During trial: Use $300 credit - Tip: Use shared-core instances (db-f1-micro) to minimize cost

BigQuery - Limit: 1TB of queries per month - Limit: 10GB storage per month - What You Can Do: - Analyze data with SQL - Practice: Datasets, tables, queries, jobs

Cloud Bigtable - Pricing: Not in always free tier

Operations¶

Cloud Monitoring - Limit: Allotment of free ingestion per month (first 50GB free per month) - What You Can Do: - Monitor resources - Practice: Metrics, alerts, dashboards

Cloud Logging - Limit: First 50GB per month free - What You Can Do: - View logs from all GCP services - Practice: Log queries, exports, sinks

Cloud Trace - Limit: First 2.5M trace spans per month free

Cloud Debugger - Free: No charge

Other Services¶

Cloud Pub/Sub - Limit: 10GB messages per month - What You Can Do: - Messaging between services - Practice: Topics, subscriptions, push/pull

Cloud Scheduler - Limit: 3 jobs per month - What You Can Do: - Cron-like job scheduling - Practice: Job scheduling, Cloud Functions triggers

Secret Manager - Limit: 6 active secret versions - Limit: 10,000 access operations per month - What You Can Do: - Store API keys, passwords - Practice: Secret management, IAM

Artifact Registry - Limit: 0.5GB storage per month - What You Can Do: - Store Docker images - Practice: Container registry

GCP Free Trial Tips¶

During 90-Day Trial: - No auto-charge after trial ends - Must explicitly upgrade to paid account - Good time to experiment with expensive services - Create GKE clusters, use Cloud SQL, test high-compute instances

After Trial: - Switch to Always Free tier - Carefully manage which services you use - Set up billing alerts

Oracle Cloud Free Tier¶

Website: https://www.oracle.com/cloud/free/

Duration: Always Free (no expiration) + $300 credit (30 days)

Credit Card Required: Yes

Why Consider: Very generous always-free tier, great for learning

Always Free Resources¶

Compute¶

Compute Instances (AMD) - Limit: 2 AMD-based Compute VMs - Specs: VM.Standard.E2.1.Micro (1 OCPU, 1GB RAM each) - What You Can Do: - Run 2 small VMs continuously - Practice: Instance management, networking

Compute Instances (ARM - Ampere A1) - Limit: 4 cores and 24GB RAM per month (Ampere A1) - What You Can Do: - Run 4 x 1-core 6GB VMs, or 2 x 2-core 12GB VMs, or 1 x 4-core 24GB VM - ARM architecture for cost-effective workloads - Practice: ARM-based computing

Note: Oracle's free ARM offering is very generous compared to competitors

Storage¶

Block Volume - Limit: 200GB total storage - What You Can Do: - Attach to compute instances - Practice: Volume management

Object Storage - Limit: 20GB - What You Can Do: - Store files, backups - Practice: Buckets, lifecycle policies

Archive Storage - Limit: 10GB - What You Can Do: - Long-term archival - Practice: Archive management

Database¶

Autonomous Database - Limit: 2 databases (20GB each) - Types: Autonomous Transaction Processing (ATP) or Autonomous Data Warehouse (ADW) - What You Can Do: - Fully managed Oracle databases - Practice: SQL, PL/SQL, database administration

NoSQL Database - Limit: 133 million reads per month - Limit: 133 million writes per month - Limit: 25GB storage per table (3 tables max)

Networking¶

Load Balancer - Limit: 1 load balancer (10Mbps bandwidth) - What You Can Do: - Load balance between VMs - Practice: Load balancing rules

Flexible Network Load Balancer - Limit: 10Mbps bandwidth

VCN (Virtual Cloud Network) - Free: Create virtual networks - What You Can Do: - Private networking - Practice: Subnets, security lists, route tables

Outbound Data Transfer - Limit: 10TB per month - Note: Much more generous than AWS (15GB) or Azure (100GB)

Other Services¶

Monitoring - Limit: 500 million ingestion datapoints - Limit: 1 billion retrieval datapoints

Notifications - Limit: 1 million sent per month

Resource Manager - Free: Infrastructure as Code (Terraform-based)

$300 Credit (30 Days)¶

• Use for any Oracle Cloud service
• Test paid services
• Expires after 30 days or when exhausted

Oracle vs Other Clouds¶

Pros: - Very generous always-free tier (especially ARM compute) - Great for learning Oracle Database - High data transfer limits - No expiration on free tier

Cons: - Smaller ecosystem than AWS/Azure/GCP - Fewer learning resources - Less relevant for most cloud certifications (except Oracle Cloud Infrastructure)

IBM Cloud Free Tier¶

Website: https://www.ibm.com/cloud/free

Duration: Lite account (always free) + $200 credit (30 days for new accounts)

Credit Card Required: No (for Lite account)

Always Free (Lite Services)¶

Compute¶

Cloud Foundry Apps - Limit: 256MB of instantaneous runtime memory - What You Can Do: - Deploy applications - Practice: PaaS deployments

Functions - Limit: 5 million executions per month - What You Can Do: - Serverless functions - Practice: OpenWhisk-based functions

Storage¶

Object Storage - Limit: 25GB per month - What You Can Do: - Store files - Practice: Buckets, access control

Database¶

Cloudant (NoSQL) - Limit: 1GB data storage - Limit: 20 queries per second - What You Can Do: - JSON document database - Practice: NoSQL queries

Db2 - Limit: 200MB storage - What You Can Do: - Relational database - Practice: SQL

AI/ML¶

Watson Assistant - Limit: 10,000 messages per month - What You Can Do: - Build chatbots - Practice: Conversational AI

Watson Studio - Limit: 50 capacity unit-hours per month - What You Can Do: - Data science projects - Practice: Machine learning

Watson Discovery - Limit: 1,000 documents per month - What You Can Do: - AI-powered search - Practice: Document analysis

Other Services¶

Kubernetes Service - Limit: 1 free cluster - What You Can Do: - Learn Kubernetes - Practice: Container orchestration

Blockchain Platform - Limit: Basic plan free - What You Can Do: - Develop blockchain applications

Why Consider IBM Cloud¶

Pros: - No credit card required for Lite account - Good AI/ML services in free tier - Unique services (Watson, Blockchain) - Great for IBM certifications

Cons: - Smaller ecosystem - Fewer learning resources - Less popular for general cloud certifications - Some free tier limits are restrictive

Cost Management Strategies¶

Set Up Billing Alerts Immediately¶

AWS CloudWatch Billing Alarms:

# Enable billing alerts in Billing Preferences first
aws cloudwatch put-metric-alarm \
  --alarm-name BillingAlert \
  --alarm-description "Alert when charges exceed $5" \
  --metric-name EstimatedCharges \
  --namespace AWS/Billing \
  --statistic Maximum \
  --period 21600 \
  --threshold 5 \
  --comparison-operator GreaterThanThreshold \
  --evaluation-periods 1

AWS Budgets: - Set up budget alerts in AWS Budgets console - Free: 2 budgets - Get alerts at 50%, 80%, 100% of budget

Azure Cost Alerts: - Set up in Cost Management + Billing - Configure budget alerts - Free: 10 budgets

GCP Billing Alerts: - Set up in Billing > Budgets & Alerts - Configure budget thresholds - Get email/Pub/Sub alerts

Use Cost Calculators Before Deployment¶

AWS Pricing Calculator: https://calculator.aws/ Azure Pricing Calculator: https://azure.microsoft.com/pricing/calculator/ GCP Pricing Calculator: https://cloud.google.com/products/calculator

Tag Everything¶

AWS:

aws ec2 create-tags --resources i-xxx --tags Key=Project,Value=CertificationStudy Key=Owner,Value=YourName

Azure:

az resource tag --tags Project=CertificationStudy Owner=YourName --resource-group myRG --name myVM --resource-type Microsoft.Compute/virtualMachines

GCP:

gcloud compute instances add-labels vm1 --labels=project=certification-study,owner=yourname

Why Tag: - Track which resources are for study - Identify cost by project - Easier cleanup

Use Spot/Preemptible/Low-Priority Instances¶

AWS Spot Instances: - Up to 90% discount - Can be interrupted - Good for: Learning, non-critical workloads

Azure Spot VMs: - Up to 90% discount - Can be evicted - Good for: Testing, learning

GCP Preemptible VMs: - Up to 80% discount - Maximum 24-hour runtime - Good for: Batch jobs, learning

Automate Shutdown¶

AWS Lambda + CloudWatch Events:

# Lambda function to stop EC2 instances
import boto3
ec2 = boto3.client('ec2')

def lambda_handler(event, context):
    instances = ['i-xxx', 'i-yyy']
    ec2.stop_instances(InstanceIds=instances)
    return 'Stopped instances: ' + str(instances)

Azure Automation: - Use Azure Automation runbooks - Schedule VM start/stop

GCP Cloud Scheduler: - Schedule instance stop with Cloud Scheduler - Use Cloud Functions to automate

Monitor Costs Daily¶

AWS: - Check AWS Cost Explorer daily - Review Cost and Usage Reports

Azure: - Check Cost Management + Billing daily - Set up cost analysis

GCP: - Check Billing Reports daily - Use BigQuery for detailed analysis

Delete Unused Resources¶

Common Forgotten Resources:

1. Elastic IPs (AWS) - Charged when not attached
2. Snapshots - Accumulate over time
3. Load Balancers - Hourly charges
4. NAT Gateways - Hourly charges
5. VPN Connections - Hourly charges
6. Stopped Instances with attached EBS - EBS still charged
7. Old AMIs and associated snapshots
8. CloudWatch Logs - Storage costs
9. S3 Buckets - Check for versioning

Weekly Cleanup Routine:

# AWS: List all resources by region
aws ec2 describe-instances --query 'Reservations[*].Instances[*].[InstanceId,State.Name,Tags[?Key==`Name`].Value|[0]]' --output table

# Check for unattached volumes
aws ec2 describe-volumes --filters "Name=status,Values=available" --output table

# Check for unassociated Elastic IPs
aws ec2 describe-addresses --query 'Addresses[?AssociationId==null]' --output table

Free Tier Best Practices¶

1. Use Free Tier Exclusively for First Month¶

Strategy: - Month 1: Only use free tier services - Learn the limits - Understand what's included - Build confidence before expanding

2. Create Separate Accounts for Different Clouds¶

Why: - Isolate free tier usage - Avoid mixing personal and learning projects - Easier to track certification study costs - Can delete entire account when done

3. Use Infrastructure as Code¶

Benefits: - Quickly tear down and rebuild environments - No forgotten resources - Reproducible setups - Version control your infrastructure

AWS CloudFormation Example:

AWSTemplateFormatVersion: '2010-09-09'
Description: Free tier EC2 instance

Resources:
  MyEC2Instance:
    Type: AWS::EC2::Instance
    Properties:
      InstanceType: t2.micro
      ImageId: ami-0c55b159cbfafe1f0
      Tags:
        - Key: Name
          Value: FreeTierStudy

Teardown:

aws cloudformation delete-stack --stack-name my-stack

4. Schedule Learning Time¶

Why: - Free tiers reset monthly - Plan major projects for start of month - More hours/requests available - Better resource management

Example Monthly Plan: - Week 1: Deploy infrastructure - Week 2: Hands-on practice - Week 3: Advanced features - Week 4: Cleanup, review

5. Leverage Multiple Free Tiers¶

Strategy: - Use AWS for EC2 and S3 practice - Use Azure for Web Apps and SQL practice - Use GCP for BigQuery and App Engine practice - Use Oracle for generous compute (ARM)

Why: - Different strengths per cloud - More free resources overall - Better multi-cloud understanding - Reduce risk of exceeding limits

6. Document Your Usage¶

Track: - What services you've used - How much quota consumed - Which projects used which resources - Lessons learned

Tools: - Spreadsheet - Notion/Obsidian - GitHub repo with notes

Multi-Cloud Strategy¶

Free Tier Comparison Table¶

Maximizing All Free Tiers Simultaneously¶

Month 1-3: GCP Focus (use $300 credit) - Deploy expensive resources (GKE, Cloud SQL) - Test high-compute instances - Learn GCP-specific services

Month 1-12: AWS Focus - Primary compute on EC2 - S3 for storage - RDS for databases - Build complex architectures

Month 1-12: Azure Focus - Web Apps for PaaS - Azure SQL for databases - Use $200 credit in first 30 days for testing

Always: Oracle Cloud - Extra compute capacity (ARM instances) - Additional storage - Oracle Database practice

Result: Massive free resources across all clouds for certification study

Study Projects Using Free Tiers¶

Project 1: Three-Tier Web Application¶

Architecture: - Frontend: Static website (S3/Blob/Cloud Storage) - Backend: EC2/VM/Compute Engine (t2.micro/B1s/e2-micro) - Database: RDS/SQL Database/Cloud SQL (free tier)

Cost: Free within limits

Time: 8-12 hours

Skills: - Compute, storage, networking, databases - Load balancing, security groups - Multi-tier architecture

Project 2: Serverless API¶

Architecture: - API: Lambda/Functions/Cloud Functions - Database: DynamoDB/Cosmos DB/Firestore - API Gateway: API Gateway/APIM/Cloud Endpoints

Cost: Free (well within free tier limits)

Time: 6-10 hours

Skills: - Serverless computing - NoSQL databases - API management

Project 3: CI/CD Pipeline¶

Architecture: - Code: GitHub/Azure Repos/Cloud Source Repositories - Build: CodeBuild/Azure Pipelines/Cloud Build - Deploy: S3/App Service/Cloud Run

Cost: Minimal (within free tier)

Time: 10-15 hours

Skills: - DevOps practices - CI/CD pipelines - Automation

Project 4: Data Analytics Pipeline¶

Architecture: - Storage: S3/Blob Storage/Cloud Storage - Processing: Lambda/Functions/Cloud Functions - Analytics: Athena/SQL Database/BigQuery

Cost: Free (within free tier)

Time: 12-18 hours

Skills: - Data engineering - Analytics - ETL processes

Common Pitfalls and How to Avoid Them¶

Pitfall 1: Data Transfer Costs¶

Problem: Data transfer out is often the largest unexpected cost

AWS: 15GB/month free Azure: 100GB/month free (12 months) GCP: 1GB/month free (Compute Engine)

How to Avoid: - Minimize data downloads - Use same-region resources - Compress data before transfer - Test on small datasets

Pitfall 2: Elastic IPs (AWS)¶

Problem: Elastic IPs cost $0.005/hour when not attached to running instance

How to Avoid: - Release IPs when not in use - Use public DNS names instead - Attach IPs only when needed

Pitfall 3: Stopped Instances with Attached Storage¶

Problem: Stopped EC2 instances still incur EBS charges

How to Avoid: - Terminate instances when done (not just stop) - Create AMIs before terminating - Use snapshots for backups, not running EBS volumes

Pitfall 4: Old Snapshots¶

Problem: Snapshots accumulate and cost $0.05/GB/month (AWS)

How to Avoid: - Delete old snapshots - Use lifecycle policies - Tag snapshots by date - Review snapshots weekly

Pitfall 5: Unused Load Balancers¶

Problem: Load balancers cost ~$0.025/hour (AWS ALB) even if not used

How to Avoid: - Delete load balancers when not needed - Use load balancers only for multi-instance scenarios - Consider alternatives (e.g., CloudFront)

Pitfall 6: NAT Gateways¶

Problem: NAT Gateways cost ~$0.045/hour (AWS) plus data processing

How to Avoid: - Use NAT instances (t2.micro) instead during learning - Delete NAT gateways when not needed - Use public subnets where possible

Pitfall 7: Forgotten Resources in Multiple Regions¶

Problem: Resources in different regions are easy to forget

How to Avoid: - Stick to one region for learning - Check all regions during cleanup - Use AWS Config or scripts to audit

AWS Check All Regions:

for region in $(aws ec2 describe-regions --query 'Regions[].RegionName' --output text); do
  echo "Region: $region"
  aws ec2 describe-instances --region $region --query 'Reservations[*].Instances[*].[InstanceId,State.Name]' --output table
done

Pitfall 8: Exceeding Free Tier by Mistake¶

Problem: Running 2 t2.micro instances uses 1,500 hours/month (750 free = 750 billable)

How to Avoid: - Run only 1 t2.micro continuously - Or run multiple instances for less time - Monitor hours used - Set up billing alerts

Pitfall 9: RDS Multi-AZ¶

Problem: Multi-AZ deployment doubles RDS costs (not in free tier)

How to Avoid: - Use single-AZ for learning - Test Multi-AZ briefly during free trial - Understand it's not in free tier

Pitfall 10: CloudWatch Logs Retention¶

Problem: CloudWatch Logs storage costs add up

How to Avoid: - Set short retention periods (1-3 days for learning) - Delete log groups when done - Export to S3 if needed long-term

Free Tier Checklist¶

Before Starting¶

• Create new account (or use existing)
• Set up MFA on root/admin account
• Add payment method (required for AWS, Azure, GCP)
• Set up billing alerts ($5, $10, $20)
• Note free tier start date
• Choose primary region

Weekly¶

• Check billing dashboard
• Review running resources
• Delete unused resources
• Check all regions for forgotten resources
• Review snapshots and backups
• Update cost tracking spreadsheet

Monthly¶

• Review entire month's costs
• Identify any unexpected charges
• Plan next month's projects
• Check free tier expiration (AWS, Azure at 12 months)

End of Certification Study¶

• Delete all resources
• Delete snapshots and AMIs
• Release Elastic IPs
• Delete S3 buckets (empty first)
• Delete CloudWatch log groups
• Deactivate account (or keep for future use)

Recommended Free Tier Learning Path¶

Month 1: Basics (AWS Focus)¶

Week 1: Setup + Core Services - Create account, set up billing alerts - Launch t2.micro EC2 instance - Create S3 bucket, upload files - Explore IAM users and roles

Week 2: Networking - Create VPC, subnets - Security groups, NACLs - Public and private subnets - NAT instances (not gateways)

Week 3: Databases - Deploy RDS MySQL (db.t2.micro) - Practice backups, snapshots - Connect from EC2

Week 4: Serverless - Create Lambda functions - DynamoDB tables - API Gateway

Month 2: Intermediate (Azure Focus)¶

Week 1: Azure Basics - Create Azure account, use $200 credit - Deploy B1S VM - Blob storage - Azure AD basics

Week 2: Azure PaaS - Deploy App Service - Azure SQL Database - Application Insights

Week 3: Azure Networking - Virtual Networks - NSGs - Load Balancer

Week 4: Azure Serverless - Azure Functions - Logic Apps - Cosmos DB

Month 3: Advanced (GCP Focus)¶

Week 1: GCP Basics (use $300 credit) - Create GCP account - Deploy e2-micro instance - Cloud Storage - IAM and Organization

Week 2: GCP PaaS - App Engine - Cloud Functions - Cloud Run

Week 3: GCP Data - BigQuery - Cloud SQL - Firestore

Week 4: GCP Kubernetes - Create GKE cluster - Deploy applications - Practice kubectl

Month 4-12: Mastery + Multi-Cloud¶

• Build complex projects across clouds
• Implement CI/CD pipelines
• Multi-cloud architectures
• Certification exam preparation
• Hands-on labs for specific certifications

Summary: Maximizing Value¶

Total Free Value (First Year): - AWS: ~$450-600 (12 months free tier) - Azure: ~$500-700 (12 months + $200 credit) - GCP: ~$600-800 ($300 credit + always free) - Oracle: ~$1200+ (always free, very generous)

Total Potential Value: $2,750-3,500 in free cloud resources

Best Strategy: 1. Start with AWS (most certifications available) 2. Add Azure and GCP in parallel 3. Use Oracle for extra compute capacity 4. Leverage IBM for AI/ML learning 5. Set up cost alerts on ALL accounts 6. Clean up religiously every week 7. Focus on free tier services 8. Use credits for expensive short-term tests

Result: Complete cloud certification study with minimal cost (mostly exam fees).

This guide provides everything you need to maximize free cloud resources for certification study. Follow the best practices, set up alerts, and clean up regularly to avoid unexpected charges.