Recommended resources
The freely available resources this lesson will direct you to.
Welcome back. In the previous lesson, you chose Regions and Availability Zones by checking hard constraints before optimizing cost. Before building more lab infrastructure, apply the same discipline to the account itself: decide what the labs may spend, then set up warnings while there is still time to act.
You have a USD $100 total lab allowance for an intensive two-week preparation period. In this lesson, you will configure an account-wide AWS cost budget and email alerts against that allowance. The result is a practical guardrail for the upcoming CloudFormation and networking labs—not a promise that AWS will automatically stop charging at a chosen amount.
Decide what the budget must protect
AWS Budgets compares costs with an amount you specify. An actual-cost alert responds to spending already recorded; a forecasted-cost alert responds when AWS projects that spending will cross a threshold during the budget period. Both are useful: an actual alert catches a real charge, while a forecast can give you time to revise a lab plan.
Watch the console walkthrough first. Its example uses a monthly budget, whereas yours will ideally cover the entire two-week lab period.
AWS Budgets Tutorial - Setup Alerts for AWS Billing | Amazon Web Services
Watch “AWS Budgets Tutorial - Setup Alerts for AWS Billing” from Amazon Web Services to see the relationship between a cost budget, its amount, and an email threshold.
Watch budget purpose for the distinction between actual and projected costs. Then watch console setup, focusing on the budget type, period, amount, and alert recipient. Console labels may differ from those in the recording; use the current AWS documentation below for the fields you configure.
A budget is not a spending cap. A charge can accrue before billing data updates or an email arrives, and an alert does not ordinarily stop EC2 instances, NAT gateways, or other resources. For a $100 total allowance, a sensible example is an $80 budget with alerts before $80. The remaining $20 is a response buffer, not a guarantee that spending cannot exceed $100. A costly misconfiguration could consume that buffer before you see an alert.
Free Tier eligibility and any credits also need separate attention. Do not treat “Free Tier account” as meaning that every service or configuration in a lab is free. For this guardrail, you want visibility into resource costs even if credits currently reduce what is due on the bill.
Create one budget for the whole lab period
The AWS Cost Management procedure documents both the period controls and the alert fields. Read it alongside the console rather than trying to memorize a sequence of screens.
Creating a cost budget - AWS Cost Management
Read “Creating a cost budget” from AWS Cost Management for the current cost-budget setup and notification options.
In “To create a cost budget,” read from the instruction to open the Billing and Cost Management console through the explanation of the budgeted amount: budget setup. Pay particular attention to Custom period, Expiring budget, and Fixed budgeting method. Later in the same procedure, find “Set alert threshold” after “Add an alert threshold”; read from its $200 example through the final review instruction: alert configuration. Notice the separate Actual and Forecasted choices and the email-verification requirement.
Now set up the budget in your account:
- Open Billing and Cost Management → Budgets → Create budget. Use Customize (advanced) and Cost budget so you can choose the period and thresholds deliberately. Sign in with an identity authorized to use Billing and Cost Management. If an IAM identity cannot open billing information, resolve the account’s billing-access settings and its IAM permissions; do not use the root user for routine lab work.
- Set the budget period to the lab window. Choose Custom, with a start date covering your first lab and an end date after your planned cleanup and final bill check. Choose Expiring budget for this one-time window, Fixed as the budgeting method, and $80 as the amount. Budget times use UTC, so allow for the date boundary if you work near midnight. If your lab continues after the end date, this budget will no longer provide ongoing coverage.
- Keep the scope account-wide. Do not filter to one service or Region: a network lab might incur compute, storage, public IPv4, data-transfer, and networking charges in several places. Review Advanced options so you understand which cost figure you are monitoring. For example, unblended costs show resource costs at the time of usage, while net unblended costs account for discounts and credits. If credits are available, monitoring only a net figure can obscure how much usage your labs are generating.
- Save only after checking the review screen. Confirm the dates, $80 amount, account-wide scope, cost option, and alert recipients. AWS Budgets needs a notification recipient or an action for each configured alert.
A monthly recurring budget is a reasonable alternative if a Custom period is unavailable in your console. But it resets each month: an $80 monthly budget does not represent an $80 total limit if your two-week course crosses a month boundary. In that case, track the combined spend across both months and set appropriate alerts for each month rather than relying on one resettable figure.
Put the alerts before the point of no return
The illustration below shows an older threshold screen. It uses a $600 budget and a forecasted threshold of 1% ($6); neither amount is appropriate to copy into your lab. Its useful feature is the distinction between the budget amount, the threshold type, and the notification recipient.
For the $80 example budget, configure these alerts using % of budgeted amount (or the equivalent absolute dollar amounts):
| Alert | Threshold | What you do |
|---|---|---|
| Actual cost | 25% ($20) | Identify which services and Regions are generating charges. |
| Actual cost | 75% ($60) | Reduce or postpone optional labs and verify cleanup of completed ones. |
| Forecasted cost | 100% ($80) | Change the plan immediately if projected spending would use the buffer. |
| Actual cost | 100% ($80) | Stop launching billable labs and remove resources no longer needed. |
For each alert, select Actual or Forecasted as shown, enter an email address you can monitor, and complete any required email verification. You can use the same recipient for all four. Confirm the budget and its alerts appear after creation; do not assume that entering an address means notifications are ready to arrive. A forecast may be less informative until AWS has enough spending history, which is why the actual-cost alerts matter too.
In SAA-C03 language, these are AWS Budgets cost alerts. A CloudWatch estimated-billing alarm is a separate possible billing-alert mechanism; neither mechanism is, by itself, an automatic resource shutdown. Budget actions exist, but adding one without understanding what it would affect is not a substitute for monitoring and cleanup in this lab account.
Use the guardrail while you build
Check the budget and Bills in Billing and Cost Management at the start of each lab day and after deploying potentially costly infrastructure. Look at the service and Region breakdown when a number rises; the AWS Budgets email tells you when to investigate, not necessarily which resource to delete. Allow for reporting delay, especially immediately after creating or removing resources.
Before a lab, estimate its likely ongoing charges. Afterward, remove what you no longer need and check again for leftovers: stopped compute may still have attached storage, and a VPC cleanup may leave other billable components if you have not inspected them. A forecast that reaches $80—or unexpectedly rapid actual spending—is a reason to pause before the next lab, even if you have not received an email.
Your configured $80 budget and alerts provide early warnings against a $100 total allowance. Keeping spending inside that allowance still depends on checking charges and dismantling resources promptly. Next, you will deploy, update, and delete a basic CloudFormation stack; this budget gives you a safer setting in which to practise that full lifecycle.