Skip to content
intermediate Phase 17 · AWS Cost Management

AWS Budgets & Alerts

Set up cost budgets and usage alerts with AWS Budgets.

45m
0 problems
Topic Progress 0%

What Budgets Covers

What Budgets Covers

AWS Budgets watches four things: cost (actual and forecast), usage (units like GB-hours), reservation utilization/coverage for RIs and Savings Plans, and custom budgets with dimensions you choose. You set a limit and thresholds — commonly 80% and 100% — and get notified by email, SNS, or Slack via Chatbot when a threshold is crossed, including forecast alerts that fire before you actually overspend.

The utilization budgets are the quiet heroes: an RI at 40% utilization or a Savings Plan drifting below target is money already committed and wasted, and nothing else will tell you. Budgets live at the account level with organization views at the payer, so each team can own its own ceiling.

Creating and Alerting

Creating and Alerting

From the console: Budgets → Create budget → Cost budget, choose monthly versus custom period, filter by service/tag/linked account (a team budget needs tag-based filtering with enforced tagging to be accurate), then add thresholds with email and SNS topics. The same flow works via aws budgets create-budget with a budget JSON and notification subscribers.

For automation, budget actions attach an IAM policy, SCP, or modify IAM permissions when a threshold trips — the difference between "we were told" and "the spend stopped." Pair with AWS Cost Anomaly Detection for unexpected spikes a static threshold would miss; budgets see the limit you imagined, anomalies see the ones you did not.

Pitfalls and Interview Focus

Pitfalls and Interview Focus

Budgets are not real-time: evaluations and notifications run periodically (roughly daily cadence), not per-charge, so they will not stop a runaway instance within the hour unless a budget action is attached — and even that runs on the same cycle. For intraday control, watch CloudWatch usage metrics or use service-native alarms.

Other traps: untagged resources escape filtered budgets (enforce tags with Config rules); free-tier resources and credits distort percentages; forecast alerts depend on historical patterns and mislead on brand-new spend; and zero-notification setups — a budget with no subscriber is decoration.

Interview: distinguish detection (budgets/anomalies) from enforcement (budget actions, SCPs), and be ready to explain why a budget alone cannot guarantee a hard cost limit.

Quiz

1. What is the key difference between a budget alert and a budget action?

Question 1 options

2. Why can a $500 cost budget miss a spike at 10 p.m.?

Question 2 options

3. Which budget type flags an underused one-year reservation?

Question 3 options

Flashcards

Question

Budget types

Answer

Cost, usage, RI/Savings Plan utilization & coverage, and custom dimensions.

Question

Budget action

Answer

Applies IAM policy/SCP or permission changes when a threshold is breached.

Question

Alert channels

Answer

Email, SNS (→ Lambda/Chatbot), Slack via Chatbot.

Question

Forecast alert

Answer

Fires when projected spend crosses the threshold before it is actually spent.

Question

Budgets vs anomalies

Answer

Budgets check limits you set; Cost Anomaly Detection finds unexpected patterns.

Revision Notes

Key Takeaways

  • 1. Budgets detect against limits; budget actions (and SCPs) enforce
  • 2. Utilization/coverage budgets catch wasted commitments, not just overspend
  • 3. Tag hygiene is a prerequisite for any meaningful team budget

Interview Tips

  • • Answer cost-control questions in layers: budgets → anomalies → actions → SCP ceiling
  • • Be explicit that budgets lag — say what you use for intraday visibility

Cheat Sheet

Budgets: cost/usage/utilization/forecast with 80%/100% thresholds → email/SNS/Chatbot. Actions attach IAM/SCP to enforce. Periodic evaluation (lags); needs enforced tags.

Continue Learning

Before You Start

  • No prerequisites listed.