AWS Service Cards

Problem it solves

Provides on-demand access to AWS compliance reports (e.g., SOC, PCI) and lets you accept agreements like the Business Associate Addendum (BAA).

When to choose it

When an auditor or compliance team needs proof of AWS's compliance certifications.

Real-world example

Downloading an ISO 27001 report from Artifact to satisfy a customer security questionnaire.

Pricing model

Free.

Problem it solves

Lets you set custom cost or usage budgets and receive alerts when actual or forecasted spend exceeds a threshold.

When to choose it

When you want proactive alerts before you overspend, rather than just reviewing historical data.

Real-world example

Getting an email alert when monthly spend is forecasted to exceed a $500 budget.

Pricing model

First two budgets are free per account; small fee per additional budget/action.

Problem it solves

Gives organizations a structured approach (across business, people, governance, platform, security, and operations perspectives) for planning a cloud migration.

When to choose it

Early in a company-wide cloud adoption or migration strategy, before individual workloads are touched.

Real-world example

A company forming a Cloud Center of Excellence uses CAF perspectives to assign workstream owners.

Pricing model

Free framework/guidance, not a billed service.

Problem it solves

Provisions and manages AWS resources using declarative "infrastructure as code" templates.

When to choose it

When you want repeatable, version-controlled infrastructure deployments instead of manually clicking through the console.

Real-world example

Defining an entire application stack (VPC, EC2, RDS, load balancer) in a YAML template that can be redeployed identically in another region.

Pricing model

Free; you pay only for the underlying resources it creates.

Problem it solves

Records every API call made in your AWS account — who did what, when, and from where — for auditing and governance.

When to choose it

When you need an audit trail of account activity, e.g. to investigate who deleted a resource or changed a security group.

Real-world example

Discovering which IAM user terminated a production EC2 instance by searching CloudTrail event history.

Pricing model

One free trail of management events per account; additional trails and data events are billed per event.

Problem it solves

Visualizes and analyzes historical AWS spending and usage patterns, and forecasts future costs.

When to choose it

When you need to understand where money is being spent, identify trends, or forecast next month's bill.

Real-world example

Discovering that a specific service's cost tripled last month due to an untagged forgotten resource.

Pricing model

Free to use.

Problem it solves

A platform-as-a-service that automatically handles deployment, capacity provisioning, load balancing, and scaling for common application stacks.

When to choose it

When you want to deploy a web app quickly without manually configuring the underlying EC2/load balancer/Auto Scaling resources.

Real-world example

Uploading a Laravel app as a zip file and letting Beanstalk provision EC2 instances, a load balancer, and Auto Scaling automatically.

Pricing model

Free; you pay only for the underlying resources (EC2, S3, etc.) it provisions.

Problem it solves

Provides single sign-on (SSO) access to multiple AWS accounts and business applications from one place.

When to choose it

When employees need one login to access several AWS accounts or SAML-enabled apps, instead of separate IAM users per account.

Real-world example

A workforce logs in once through a company identity provider and sees a portal listing every AWS account they can access.

Pricing model

Free.

Problem it solves

Controls who (users, groups, roles) can do what (permissions) on which AWS resources.

When to choose it

Anytime you need to grant a person, application, or service least-privilege access to AWS resources.

Real-world example

Creating a role that lets an EC2 instance read from a specific S3 bucket without embedding access keys.

Pricing model

Free.

Problem it solves

Creates and controls encryption keys used to encrypt data across AWS services.

When to choose it

Whenever you need to encrypt data at rest (S3, EBS, RDS, etc.) and control/rotate/audit who can use the key.

Real-world example

Encrypting an RDS database with a customer-managed key so only specific IAM roles can decrypt it.

Pricing model

Monthly fee per key, plus per-request charges.

Problem it solves

Runs your code in response to events without provisioning or managing any servers.

When to choose it

Short-lived, event-driven tasks (e.g., resizing an image on upload, an API backend) where you don't want to manage infrastructure.

Real-world example

Automatically generating a thumbnail whenever a new image is uploaded to S3.

Pricing model

Pay per request and per GB-second of compute time used; generous free tier.

Problem it solves

Centrally manages and governs multiple AWS accounts, including consolidated billing and account-wide guardrails.

When to choose it

When a company has (or is growing into) more than one AWS account and needs consistent policy enforcement and one consolidated bill.

Real-world example

Using an SCP to block every account in the "Sandbox" OU from launching expensive instance types.

Pricing model

Free.

Problem it solves

Estimates the cost of AWS services before you provision them, based on configuration and usage assumptions.

When to choose it

During planning/design, to compare architecture options and estimate monthly costs before committing.

Real-world example

Estimating the monthly cost difference between running a workload on EC2 On-Demand versus Fargate before choosing an architecture.

Pricing model

Free to use.

Problem it solves

Protects applications running on AWS against Distributed Denial of Service (DDoS) attacks.

When to choose it

For any internet-facing application; Standard is automatic and free, Advanced adds cost protection and 24/7 DDoS response support.

Real-world example

A public website automatically absorbing a volumetric traffic flood via Shield Standard with no manual action.

Pricing model

Standard is free and automatic; Advanced is a paid subscription with commitment.

Problem it solves

Tiered support plans (Basic, Developer, Business, Enterprise On-Ramp, Enterprise) providing different levels of technical support and response times.

When to choose it

Choose based on business criticality: production workloads generally need at least Business support for faster response times and full Trusted Advisor checks.

Real-world example

A startup upgrading to Business Support to get 24/7 access to Cloud Support Engineers with under-1-hour response for production-down issues.

Pricing model

Basic is free; other tiers are a percentage of monthly AWS usage or a flat fee, depending on plan.

Problem it solves

Inspects your live AWS account and gives real-time recommendations across cost optimization, security, fault tolerance, performance, and service limits.

When to choose it

Ongoing account health checks — e.g., finding idle resources, open security groups, or approaching service limits.

Real-world example

Getting flagged that an Elastic IP address is unattached and quietly incurring charges.

Pricing model

Core checks are free for all accounts; full check list requires Business or Enterprise Support.

Problem it solves

Filters and blocks malicious HTTP/HTTPS requests, such as SQL injection and cross-site scripting, before they reach your application.

When to choose it

When you need layer-7 (application layer) protection for APIs, CloudFront distributions, or load balancers.

Real-world example

Blocking requests that contain SQL injection patterns before they reach a web app behind an Application Load Balancer.

Pricing model

Pay per web ACL, rule, and request processed.

Problem it solves

Reviews a workload against six pillars (operational excellence, security, reliability, performance efficiency, cost optimization, sustainability) and surfaces risks.

When to choose it

Before or after building a workload, when you need a structured way to identify architectural risks and get improvement recommendations.

Real-world example

Running a workload review before a major product launch to catch single points of failure.

Pricing model

Free to use.

Problem it solves

A content delivery network (CDN) that caches content at edge locations close to users, reducing latency.

When to choose it

Serving static or dynamic content globally with low latency, or reducing load on an origin server (like S3 or EC2).

Real-world example

Caching a website's images and JS bundles at edge locations so users worldwide get fast load times.

Pricing model

Pay per GB of data transferred out and number of HTTP/HTTPS requests, varying by edge location.

Problem it solves

Collects metrics, logs, and alarms to monitor the performance and health of AWS resources and applications.

When to choose it

When you need dashboards, alerting on thresholds (like high CPU), or centralized application logs.

Real-world example

Setting an alarm that triggers Auto Scaling to add EC2 instances when average CPU exceeds 70%.

Pricing model

Free tier for basic metrics; pay for custom metrics, dashboards, alarms, and log storage/ingestion.

Problem it solves

A fully managed, serverless NoSQL database delivering single-digit-millisecond performance at virtually any scale.

When to choose it

High-throughput key-value or document workloads (e.g., session state, shopping carts) where schema flexibility and horizontal scale matter more than SQL joins.

Real-world example

Storing millions of user session records for a mobile game with unpredictable, spiky traffic.

Pricing model

Pay per read/write request unit (on-demand) or provisioned capacity, plus storage.

Problem it solves

Provides persistent block storage volumes that attach to a single EC2 instance, like a virtual hard drive.

When to choose it

When an EC2 instance needs low-latency, persistent storage for a database, boot volume, or application data.

Real-world example

The root volume of an EC2 instance running a MySQL database.

Pricing model

Pay per GB provisioned per month, plus IOPS/throughput for certain volume types.

Problem it solves

Provides resizable virtual servers (instances) you fully control, for workloads that need custom OS/runtime configuration.

When to choose it

Long-running applications, legacy software, or when you need control over the OS and installed software.

Real-world example

Hosting a custom Laravel application that needs specific PHP extensions and persistent background workers.

Pricing model

On-Demand (per second/hour), Reserved (commitment discount), Spot (spare capacity discount), or Savings Plans.

Problem it solves

Runs and orchestrates Docker containers; Fargate removes the need to provision or manage the underlying EC2 instances.

When to choose it

When your app is already containerized and needs more control/duration than Lambda but less operational overhead than raw EC2.

Real-world example

Running a containerized microservices API where each service scales independently.

Pricing model

ECS on EC2: pay for underlying EC2. Fargate: pay per vCPU/memory reserved per second.

Problem it solves

Provides a scalable, shared network file system that many EC2 instances or containers can mount at the same time.

When to choose it

When multiple compute resources need concurrent read/write access to the same shared files.

Real-world example

A fleet of web servers behind a load balancer all reading and writing to the same uploaded-files directory.

Pricing model

Pay per GB stored, with pricing tiers for storage class and throughput mode.

Problem it solves

Continuously monitors accounts and workloads for malicious activity using threat intelligence and machine learning.

When to choose it

When you want automated, always-on threat detection without managing your own intrusion-detection infrastructure.

Real-world example

Getting alerted when an EC2 instance starts communicating with a known cryptocurrency-mining domain.

Pricing model

Pay based on volume of logs and events analyzed.

Problem it solves

Manages relational databases (MySQL, PostgreSQL, etc.), automating backups, patching, and failover.

When to choose it

Structured, relational data with well-defined schemas and the need for SQL joins/transactions, without wanting to manage the database server yourself.

Real-world example

Storing customer orders and inventory in a normalized relational schema for an e-commerce app.

Pricing model

Pay per instance-hour, storage provisioned, and I/O, based on engine and instance class.

Problem it solves

A highly available Domain Name System (DNS) service that translates domain names into IP addresses and can also register domains.

When to choose it

Whenever you need to point a domain name at AWS resources, or need advanced routing like latency-based or health-check failover routing.

Real-world example

Configuring failover DNS routing so traffic automatically shifts to a backup region if the primary region's health check fails.

Pricing model

Pay per hosted zone per month, plus per DNS query and domain registration fees.

Problem it solves

Stores and retrieves any amount of object data (files) with high durability, accessible over the internet via API.

When to choose it

Static website assets, backups, data lakes, or any file storage that doesn't need a traditional filesystem or block device.

Real-world example

Hosting a static site's images and video files served through CloudFront.

Pricing model

Pay per GB stored (varies by storage class), plus requests and data transfer out.

Problem it solves

A pub/sub messaging service that pushes messages to many subscribers (email, SMS, Lambda, SQS queues) at once.

When to choose it

Fan-out notifications, e.g. alerting multiple downstream systems the instant an event happens.

Real-world example

Sending an order-confirmation event to email, an SMS gateway, and a fulfillment Lambda function simultaneously.

Pricing model

Pay per number of notifications published/delivered.

Problem it solves

A fully managed message queue that decouples producers and consumers, holding messages until they're processed.

When to choose it

Decoupling application components so a spike in incoming work doesn't overwhelm downstream processing.

Real-world example

Buffering incoming image-processing jobs so a fleet of workers can pull and process them without dropping requests during traffic spikes.

Pricing model

Pay per million requests, with a monthly free tier.

Problem it solves

Lets you provision a logically isolated, private network within AWS where you control IP ranges, subnets, and routing.

When to choose it

Essentially every workload — it's the private network foundation your EC2 instances, RDS databases, and other resources live in.

Real-world example

Placing a database in a private subnet with no direct internet route, while web servers sit in a public subnet.

Pricing model

The VPC itself is free; you pay for components like NAT gateways, data transfer, and VPN connections.