Chapter 4

Billing, Pricing, and Support

AWS Certified Cloud Practitioner (CLF-C02) Study GuideBased on the AWS CLF-C02 exam objectives

The Money Chapter: Why AWS Pricing Exists the Way It Does

Here is a confession most cloud trainers will not make on page one: the billing domain of the Cloud Practitioner exam is where people fail. Not because the material is hard — it isn't, there is no math more complicated than percentages — but because it is boring, so candidates skim it, and then the exam hits them with five questions about Savings Plans versus Reserved Instances and they lose points they could have had for free.

So let us make a deal. This chapter is not boring. It is the chapter about how Amazon makes its money, and once you understand why AWS prices things the way it does, the entire domain — pricing models, cost tools, support plans — collapses into one coherent story. The exam tests this domain at roughly 18 percent of your score, which means it is worth nearly a fifth of your pass mark. Read it properly and it is free marks.

The story starts with a datacenter. Amazon owns millions of servers, and servers cost money whether they are running or not. The building, the cooling, the fiber to the building, the staff — these are sunk costs. Every hour a server sits idle, Amazon burns money. So AWS's entire pricing philosophy is built around one question: how do we get customers to commit to using capacity in advance, so our utilization stays high?

Everything in this chapter is an answer to that question. Pay-as-you-go is the default because it is flexible and it gets you in the door. Reserved Instances and Savings Plans are discounts AWS gives you in exchange for commitment — they are literally AWS paying you to help them plan capacity. Spot Instances are AWS selling you the leftover, unsold capacity at a massive discount, because 60 percent of something beats 100 percent of nothing. Once you see this logic, you will never confuse the pricing models again.

The Four Pricing Models (and the One Idea Behind Them All)

Pay-as-you-go: the default

Pay-as-you-go means exactly what it says. You spin up an EC2 instance, the meter starts running. You shut it down, the meter stops. You are billed per second (with a 60-second minimum) for EC2, per request for Lambda, per gigabyte-month for S3. There is no commitment, no upfront payment, no minimum.

This is the model every new AWS customer starts with, and it is deliberately the most expensive per unit of compute. Why? Because you are buying optionality — the right to scale up and down with zero friction. AWS charges a premium for that freedom, and then offers you discounts the moment you give some of it up.

Reserved Instances: commitment in exchange for discount

A Reserved Instance (RI) is a billing discount applied to on-demand usage in exchange for a one- or three-year commitment to a specific instance type in a specific region. Discounts run up to 72 percent versus on-demand pricing. You do not get a physical server reserved for you — despite the name, nothing is actually "reserved." It is a billing construct: AWS looks at your on-demand usage, sees that it matches the RI you purchased, and applies the discount.

There are two flavors, and the exam distinguishes them:

Standard RIs give the biggest discount (up to 72%) but lock you to the instance family. If you buy a Standard RI for m5.large instances, you cannot change it to c5.large later. You can sell it on the Reserved Instance Marketplace, but you cannot modify the family.

Convertible RIs give a smaller discount (up to 54%) but let you change the instance family, operating system, and tenancy during the term, as long as the total value stays the same. Flexibility costs money — that is the trade, every time.

You also choose a payment option: all upfront (biggest discount), partial upfront, or no upfront (smallest discount, but zero cash outlay). More money now, more discount later — the same logic as buying in bulk at a warehouse store.

Spot Instances: the bargain basement

Spot Instances let you bid on unused EC2 capacity at discounts of up to 90 percent. The catch — and it is a real one — is that AWS can reclaim the instance with a two-minute warning whenever it needs the capacity back. Your workload gets interrupted. No negotiation, no appeal.

So Spot is perfect for workloads that are fault-tolerant, stateless, and interruptible: batch processing jobs, big data analytics, CI/CD build pipelines, image rendering, anything where losing a worker mid-task just means the job retries. It is catastrophically wrong for databases, production web servers, or anything stateful where an interruption means data loss.

Think of it like standby airline tickets. You fly for almost nothing, but you only get a seat if the plane is not full, and you can be bumped. Nobody books a standby ticket for their wedding day. Same logic.

Savings Plans: the modern answer

Savings Plans are AWS's newer, more flexible commitment model. Instead of committing to a specific instance type like an RI, you commit to a dollar amount of compute usage per hour for one or three years — say, $10/hour — and in exchange you get discounted rates on whatever compute you use, up to that commitment.

There are three types, and the exam tests the distinction:

Compute Savings Plans apply to EC2, Fargate, and Lambda, across regions, instance families, and operating systems. Maximum flexibility, discount up to 66 percent.

EC2 Instance Savings Plans apply only to EC2, but lock to an instance family in a region. Less flexible, bigger discount — up to 72 percent, same as Standard RIs.

SageMaker Savings Plans apply to SageMaker machine learning usage. Discount up to 64 percent.

The exam-level summary: Savings Plans are like RIs but measured in dollars per hour instead of specific instances. If a question emphasizes flexibility across instance types and regions, the answer is Savings Plans (specifically Compute Savings Plans). If it emphasizes maximum discount with no need to change, the answer is Standard RIs or EC2 Instance Savings Plans.

Putting the four models side by side

The Tools: Seeing and Controlling Your Bill

AWS gives you four tools for understanding and managing cost, and the exam wants you to know which tool does what. They form a pipeline: estimate before you build, monitor while you run, alert before it hurts, analyze after the fact.

AWS Pricing Calculator: estimate before you spend

The AWS Pricing Calculator (which replaced the older Simple Monthly Calculator) lets you model the cost of an architecture before you deploy it. You pick services, enter usage estimates — so many EC2 hours, so much S3 storage, so much data transfer — and it produces a cost estimate you can share with your team or your boss.

The exam distinction that matters: the Pricing Calculator is for estimating future costs of planned architectures. It does not show you what you have already spent. That is a different tool.

AWS Cost Explorer: analyze what you spent

Cost Explorer is the rear-view mirror. It visualizes your historical AWS spending — by service, by region, by tag, by day — and lets you slice the data to find out where the money went. It also provides forecasting: based on your past usage, it predicts what you will spend next month.

Default retention is 12 months of data, and you can generate custom reports. The key exam phrase is "visualize, understand, and manage AWS costs over time." Over time — past and projected. That is Cost Explorer.

AWS Budgets: alert before it hurts

AWS Budgets lets you set thresholds — cost budgets ("alert me when we exceed $500 this month"), usage budgets, RI utilization budgets, Savings Plans budgets — and sends notifications via email or SNS when you cross them, or are forecast to cross them.

The critical distinction from Cost Explorer: Budgets is proactive. It watches and warns. Cost Explorer is reactive. It shows and explains. The exam will test this exact split.

The Billing Dashboard: the front door

The AWS Billing Dashboard (sometimes called the Billing and Cost Management console) is the central place where you see your current bill, past invoices, payment history, and links to all the tools above. If a question asks "where do you view your AWS bill," this is the answer. Simple, but the exam asks it.

Cost Anomaly Detection (bonus awareness)

AWS Cost Anomaly Detection uses machine learning to spot unusual spending patterns and alert you — the $3,000 spike on a Saturday night when someone left a GPU instance running. The exam mentions it lightly; know that it exists and that it is automatic detection of unusual spend, distinct from Budgets' threshold alerts.

Paying Less as an Organization

Everything so far assumed one AWS account. Real companies have dozens. AWS Organizations and Consolidated Billing exist to manage cost across all of them — and the exam treats this as a billing topic, not just a management one, because the financial benefits are the point.

AWS Organizations: one bill, many accounts

AWS Organizations lets you centrally manage multiple AWS accounts. You create an organization, invite accounts (or create new ones), and arrange them into Organizational Units (OUs) — groupings like "Production," "Development," "Finance" that mirror how your company is structured.

The billing superpower is consolidated billing: all member accounts' usage rolls up into a single bill paid by the management account. Why does this matter financially? Two reasons the exam cares about:

1. Volume discounts aggregate. S3 pricing gets cheaper per gigabyte as you store more. With consolidated billing, ten accounts each storing 10 TB get priced as 100 TB — hitting the cheaper tier. Separately, each pays the expensive tier.
2. Reserved Instances and Savings Plans share. An unused RI discount in the development account automatically applies to matching on-demand usage in the production account. Commitment discounts float across the organization to wherever they are needed.

Cost allocation tags: knowing who spent what

Consolidated billing tells you the total. Cost allocation tags tell you whose total it is. You tag resources with key-value pairs — CostCenter: Marketing, Project: Phoenix, Environment: Production — then activate the tags in the Billing console, and Cost Explorer breaks spending down by tag.

Without tags, a consolidated bill is one giant number and nobody knows which team blew the budget. With tags, the finance team can charge each department for exactly what it used. The exam phrase is "track and allocate costs" or "identify which department incurred which costs."

The TCO Conversation: Cloud Versus the Server Room

TCO — Total Cost of Ownership — is the exam's way of asking: is the cloud actually cheaper than owning hardware? The honest answer is "it depends," and AWS provides the AWS Migration Evaluator (formerly TCO Calculator) to model it.

The comparison is never just "server price versus EC2 price." True TCO includes the obvious costs (hardware, software licenses) and the hidden ones the exam loves to list: datacenter floor space, power and cooling, network hardware, IT staff salaries, maintenance contracts, and the cost of over-provisioning for peak capacity you use twice a year.

The cloud's financial argument rests on three ideas:

1. No upfront CapEx. You do not buy servers; you rent them by the second. Capital expenditure becomes operational expenditure, which matters enormously to company cash flow.
2. No over-provisioning. On-premises, you buy for the peak — the Black Friday traffic spike — and that capacity sits idle the other 364 days. In the cloud, elasticity means you scale to the peak for hours, then scale back. You stop paying for idle.
3. Undifferentiated heavy lifting disappears. Patching, racking, cooling, replacing failed drives — AWS does it, and the cost is spread across millions of customers.

AWS Support Plans: Paying for Help

AWS Support comes in five tiers, and the exam tests them as a ladder — each rung adds more, and the questions almost always hinge on response time or specific features like the TAM or Infrastructure Event Management.

Basic: free, and deliberately thin

Every AWS account gets Basic support for free. You get 24/7 access to customer service for account and billing issues, access to documentation and whitepapers, and the AWS Personal Health Dashboard. What you do not get: any technical support. If your EC2 instance will not boot, Basic support will not help you fix it.

Developer: for experimenting

Developer support (starting around $29/month) adds business-hours access to Cloud Support Associates via email, with response times under 24 hours for general guidance and under 12 hours for system-impaired issues. It is aimed at people learning and prototyping — one person experimenting, not a production workload.

Key exam detail: Developer support has no phone support and no sub-one-hour response. If the scenario needs either, Developer is wrong.

Business: for production workloads

Business support (starting around $100/month) is the first tier suitable for production. It adds 24/7 phone and chat access to cloud support engineers, response times under one hour for production-system-down emergencies, IEM — Infrastructure Event Management — for planned events like launches and migrations, and access to the full set of Trusted Advisor checks (more on Trusted Advisor in a moment).

Enterprise On-Ramp and Enterprise: for serious scale

Enterprise On-Ramp (starting around $5,500/month) adds a pool of Technical Account Managers (TAMs), cost optimization guidance, and a 30-minute response time for business-critical issues. Think of it as Enterprise-lite for companies growing into serious AWS usage.

Enterprise (starting around $15,000/month) is the top tier: a dedicated TAM (not a pool — your own named human), 15-minute response for business-critical system-down events, the Concierge Support Team for billing and account issues, and proactive guidance including Well-Architected reviews.

The TAM distinction is the exam's favorite: a pool of TAMs is Enterprise On-Ramp; a dedicated TAM is Enterprise. One word changes the answer.

AWS Trusted Advisor: the free consultant

Trusted Advisor is an automated tool that inspects your AWS environment and flags opportunities across five categories: cost optimization (idle resources, underutilized instances), performance, security, fault tolerance, and service limits. Basic support gets only the core security and service-limit checks; Business and Enterprise unlock all of them.

The exam treats Trusted Advisor as a best-practices checker, not a billing tool — though cost optimization is one of its categories. If a question asks "which service provides real-time guidance on cost optimization and security best practices," the answer is Trusted Advisor.

The Well-Architected Framework: Six Pillars, One Philosophy

We close with the AWS Well-Architected Framework — not a service but a philosophy, a set of six pillars that describe what a good cloud architecture looks like. The CLF-C02 exam tests this at a conceptual level: know the six pillar names and, roughly, what each one is about.

1. Operational Excellence — run and monitor systems to deliver business value. Think: automation, documentation, runbooks, learning from failure. The "how we operate" pillar.
2. Security — protect data, systems, and assets. Think: least privilege, encryption, detective controls. (This maps directly to the shared responsibility model from Chapter 2 of most courses.)
3. Reliability — recover from failure and keep working. Think: backups, multi-AZ deployments, auto-scaling, chaos testing. The "it keeps running" pillar.
4. Performance Efficiency — use resources efficiently as demand changes. Think: right-sizing, serverless, experimenting with new instance types. The "no waste, no lag" pillar.
5. Cost Optimization — avoid unnecessary costs. Think: pay-as-you-go discipline, Reserved Instances, turning things off, Cost Explorer. This entire chapter lives under this pillar.
6. Sustainability — minimize environmental impact. Think: efficient instance types, managed services, matching supply to demand so no energy is wasted. The newest pillar, added in 2021.

The Framework also gives us the perfect closing thought for this chapter. Notice how the pillars interact: Reliability without Cost Optimization means three copies of everything running at all times — reliable, but ruinously expensive. Cost Optimization without Reliability means one fragile server — cheap until it dies. The Framework's real lesson is that good architecture is balanced architecture, and every pricing decision you make is a small vote in that balance.

Chapter Summary

• AWS pricing is built on one idea: discounts in exchange for commitment (RIs, Savings Plans) or for taking risk (Spot).
• On-demand = maximum flexibility, no commitment. Standard RIs = maximum discount, no flexibility. Convertible RIs = middle ground. Spot = up to 90% off, interruptible. Compute Savings Plans = flexible commitment in $/hour.
• Pricing Calculator estimates future costs; Cost Explorer analyzes past costs; Budgets alerts on thresholds; Migration Evaluator compares cloud vs. on-premises TCO.
• Consolidated billing (via AWS Organizations) aggregates volume discounts and shares RI/Savings Plan discounts across accounts. Cost allocation tags attribute spend to teams and projects.
• Support ladder: Basic (free, billing only) → Developer (~$29, business-hours email) → Business (~$100, 24/7 phone, <1 hr critical, IEM) → Enterprise On-Ramp (~$5,500, TAM pool, <30 min) → Enterprise (~$15,000, dedicated TAM, <15 min).
• Trusted Advisor checks your environment against best practices in five categories.
• The Well-Architected Framework has six pillars: Operational Excellence, Security, Reliability, Performance Efficiency, Cost Optimization, Sustainability.

End-of-Chapter Test

1. A startup runs a web application with highly unpredictable traffic. There are traffic spikes several times a day, but the baseline load is low. Which EC2 pricing model is MOST appropriate?

A. Standard Reserved Instances with a three-year term B. On-demand Instances C. Spot Instances for the web servers D. Convertible Reserved Instances with a one-year term

2. A company has a steady-state database workload that will run unchanged for the next three years. They want the maximum possible discount on EC2 and will not change the instance type. Which option should they choose?

A. Compute Savings Plans with a one-year term B. Spot Instances C. Standard Reserved Instances with a three-year term, all upfront D. On-demand Instances with AWS Budgets

3. A media company needs to transcode thousands of video files overnight. The workload is fault-tolerant and can be interrupted and resumed without data loss. Which pricing model offers the LOWEST cost?

A. On-demand Instances B. Standard Reserved Instances C. Spot Instances D. EC2 Instance Savings Plans

4. A solutions architect wants to estimate the monthly cost of a planned three-tier web application BEFORE deploying any resources. Which AWS tool should be used?

A. AWS Cost Explorer B. AWS Budgets C. AWS Pricing Calculator D. AWS Cost Anomaly Detection

5. A company wants to receive an email notification when its AWS spending is forecast to exceed $2,000 in the current month. Which service meets this requirement?

A. AWS Cost Explorer B. AWS Budgets C. AWS Pricing Calculator D. AWS Trusted Advisor

6. An enterprise has 40 AWS accounts and wants a single bill, aggregated volume discounts on S3 storage, and automatic sharing of unused Reserved Instance discounts across accounts. Which AWS feature provides these benefits?

A. AWS Cost Explorer with consolidated views B. Consolidated billing through AWS Organizations C. Cost allocation tags D. AWS Budgets with linked accounts

7. A finance team needs to determine exactly how much each department (Engineering, Marketing, Finance) spends on AWS each month. What should they implement?

A. Separate AWS Organizations for each department B. Cost allocation tags activated in the Billing console C. AWS Budgets per department D. Individual support plans per department

8. A company runs a production e-commerce website and requires 24/7 phone support with a response time of under one hour for critical system failures. What is the MINIMUM AWS Support plan that meets these requirements?

A. Basic B. Developer C. Business D. Enterprise

9. Which AWS Support plan provides a DEDICATED Technical Account Manager (TAM) and a 15-minute response time for business-critical system-down events?

A. Business B. Enterprise On-Ramp C. Enterprise D. Developer

10. A company is designing a new application and wants to follow AWS best practices for recovering automatically from infrastructure failures. Which pillar of the AWS Well-Architected Framework does this concern?

A. Operational Excellence B. Security C. Reliability D. Sustainability

Answer Key

1. B — Unpredictable, spiky traffic with a low baseline means no commitment is sensible. On-demand (pay-as-you-go) lets the company scale up and down freely. Reserved Instances (A, D) require predictable steady-state usage. Spot Instances (C) are wrong because web servers serving live traffic cannot tolerate interruption.

2. C — "Maximum discount" + "will not change the instance type" + "three years" points directly to Standard Reserved Instances, three-year term, all upfront — the deepest discount AWS offers on EC2. Compute Savings Plans (A) are more flexible but discount less. Spot (B) is interruptible and unsuitable for a database. Budgets (D) only alert on spending; they do not reduce it.

3. C — Fault-tolerant, interruptible batch work is the textbook Spot Instance use case, with discounts up to 90% — the lowest cost of any option. On-demand (A) and RIs/Savings Plans (B, D) all cost more per hour for this workload profile.

4. C — The Pricing Calculator estimates the cost of planned architectures before deployment. Cost Explorer (A) analyzes historical spend. Budgets (B) set alerts on thresholds. Cost Anomaly Detection (D) flags unusual spending automatically — none of these estimate future costs of unbuilt architectures.

5. B — AWS Budgets sends notifications (including forecast-based alerts) when spending crosses or is projected to cross a threshold. Cost Explorer (A) visualizes past spend but does not send alerts. The Pricing Calculator (C) estimates future architectures. Trusted Advisor (D) checks best practices, not spending thresholds.

6. B — Consolidated billing through AWS Organizations rolls all member accounts into one bill, aggregates usage for volume discount tiers, and automatically shares unused RI and Savings Plan discounts across accounts. Cost Explorer (A) only visualizes; tags (C) attribute costs but do not consolidate billing; Budgets (D) only alert.

7. B — Cost allocation tags (key-value pairs like Department: Engineering), once activated in the Billing console, let Cost Explorer break spending down by tag — exactly what chargeback/showback requires. Separate Organizations (A) is overkill and loses consolidated-billing benefits. Budgets (C) alert but do not attribute. Support plans (D) are unrelated to cost attribution.

8. C — Business is the minimum tier with 24/7 phone/chat support and sub-one-hour response for production-system-down events. Basic (A) offers no technical support. Developer (B) is business-hours email only with slower response times. Enterprise (D) meets the requirement but is not the minimum.

9. C — Enterprise provides a dedicated TAM and 15-minute response for business-critical events. Enterprise On-Ramp (B) provides only a pool of TAMs and 30-minute response. Business (A) has no TAM. Developer (D) has no TAM and much slower response.

10. C — Automatic recovery from infrastructure failure is the definition of the Reliability pillar. Operational Excellence (A) covers running and monitoring processes. Security (B) covers protection of data and systems. Sustainability (D) covers minimizing environmental impact.