Cloud Pricing Basics — Pay-as-You-Go Model

Direct Answer

Cloud bills combine resource quantity, duration, request counts, provisioned capacity, storage, operations, support, licenses, and network transfer according to each service's price model. Ingress is often free but not universally, and egress/region/AZ rules vary. Developers influence cost through architecture, data movement, retention, right-sizing, scaling limits, tags, budgets, and deletion of unused resources.

Code Example

// Spring Boot configuration with resource tags for cost tracking
@Configuration
public class AwsResourceConfig {
    
    @Bean
    public AmazonEC2 ec2Client() {
        return AmazonEC2ClientBuilder.standard()
            .withRegion(Regions.US_EAST_1)
            .build();
    }
    
    // Tag resources with team and environment for cost allocation
    public void tagInstance(String instanceId, String team, String env) {
        var ec2 = ec2Client();
        
        var tags = List.of(
            new Tag("Team", team),        // Track which team owns this cost
            new Tag("Environment", env),  // Identify dev vs prod
            new Tag("AutoShutdown", env.equals("dev") ? "yes" : "no")
        );
        
        var request = new CreateTagsRequest()
            .withResources(instanceId)
            .withTags(tags);
        
        ec2.createTags(request);  // AWS Cost Explorer groups by these tags
    }
}

Common Mistakes & Pitfalls

• Leaving dev resources running 24/7: Freshers spin up an EC2 instance for testing, forget about it, and get a 200-dollar bill a month later when the instance sat idle for 90 percent of the time—always tag resources and set up auto-shutdown.
• Over-provisioning for fear of slowness: choosing a db.m5.4xlarge RDS instance (16 vCPU, 64 GB RAM, 320 dollars per month) for a personal project because it sounds fast, when a db.t3.micro (2 vCPU, 1 GB RAM, 15 dollars per month) would handle the actual 50 requests per day load.
• Ignoring data transfer costs: building an API that returns entire database tables as JSON responses without pagination, then wondering why the AWS bill shows 100 dollars in data transfer charges—egress is expensive, always paginate large result sets.