Pricing, billing and cost control

Why cost is an engineering skill

In a data centre, the hardware is already paid for, so an idle server costs nothing extra. In the cloud, every running resource is a meter that keeps ticking. Cloud engineers are expected to know roughly what their design will cost and to notice when the bill jumps. Some companies call this FinOps: engineers, finance and managers sharing responsibility for cloud spend.

The three main cost drivers

• Compute: servers, containers and functions, billed by the second, hour or request. An EC2 instance is billed while it is running.
• Storage: gigabytes stored per month (GB-month), for example in S3 or on EBS disks. A stopped EC2 instance still pays for its disk.
• Data transfer: data coming IN from the internet is usually free. Data going OUT to the internet is charged per GB, after a free monthly allowance (AWS currently includes 100 GB a month of data out to the internet across most services; check the pricing page). Data between availability zones or regions usually costs a little too.

Many services add other meters: requests (S3 GET and PUT calls), public IPv4 addresses (charged by the hour), NAT gateways (by the hour and by the GB), and snapshots. Always read the pricing page of a service before you use it at scale.

Pricing models for compute

A commitment is charged whether you use it or not. If your server runs only 200 hours a month, an hourly commitment for all 730 hours can cost more than On-Demand. The rule of thumb: commit only to the part of your usage that runs all the time, and leave the spiky part On-Demand.

Tools that keep the bill under control

• AWS Pricing Calculator: estimate a design before you build it, and share the estimate as a link.
• Cost Explorer: charts of what you spent, by service, region, account or tag, with a forecast.
• AWS Budgets: set a monthly amount and get emails (or other alerts) at thresholds such as 80% of actual or 100% of forecast.
• Cost Anomaly Detection: warns you when spending suddenly behaves unusually.
• Cost allocation tags: labels such as Project=results-portal or Environment=dev. After you activate them in the Billing console, Cost Explorer can split the bill by them.
• Free Tier usage alerts and the Free Tier page in the Billing console show how much of your free offers you have used.

aws ec2 create-tags --resources i-0abc1234def567890 \
    --tags Key=Project,Value=results-portal Key=Environment,Value=dev

aws ce get-cost-and-usage \
    --time-period Start=2026-10-01,End=2026-10-19 \
    --granularity MONTHLY --metrics UnblendedCost

A worked example

Say a small server costs about US$0.01 an hour, storage costs about US$0.08 per GB-month and data out costs about US$0.09 per GB after 100 free GB. (These are rounded, illustrative rates; check the current pricing page for your region.) Running the server for a full month of 730 hours costs 730 × 0.01 = US$7.30. A 30 GB disk costs 30 × 0.08 = US$2.40. Sending 150 GB to users means 50 chargeable GB, 50 × 0.09 = US$4.50. Total: about US$14.20 a month, before tax.

Tax on Indian bills

Indian customers are billed with GST (Goods and Services Tax) added to their AWS charges. Accounts with an Indian address are usually served by Amazon Web Services India Private Limited (AISPL), which bills in rupees. Check your invoice for the exact tax lines, and give your GSTIN in the tax settings if your company can claim input tax credit.