AWS FinOps Tools: The Complete Free Native Stack
You can run a real FinOps practice on AWS native tools alone. Cost Explorer gives you the interactive view, Cost and Usage Reports give you the queryable truth, Budgets and Cost Anomaly Detection give you the alerts, and Compute Optimizer gives you the right-sizing candidates. Almost all of it is free, with three specific exceptions that catch people out. Third-party tooling earns its price later, when you need chargeback across many teams or continuous automated action, not on day one.
Native tools that cost nothing
Most of them
Ones that quietly bill you
3
Time to a working baseline
30 days
SEO Focus Topics
Key Takeaways
- • The native stack is genuinely enough for a first FinOps pass. Cost Explorer, CUR, Budgets, Cost Anomaly Detection, Cost Categories and Compute Optimizer cover visibility, alerting and right-sizing.
- • Three things quietly cost money: the Cost Explorer API bills per request while the console does not, AWS Budgets bills per budget per day beyond a small free allowance, and the Trusted Advisor cost checks are gated behind a paid support plan.
- • Cost Explorer is for answering questions in minutes. CUR queried through Athena is for answering questions the console cannot express. You want both, and they are not substitutes.
- • Group CUR by usage type rather than by service, or you will never find the data transfer and NAT Gateway spend hiding inside EC2-Other.
- • The native stack runs out at three specific points: cross-team chargeback, continuous automated right-sizing, and unified reporting across more than one cloud. That is when third-party tooling starts paying for itself.
- • Tooling is not the hard part. A monthly review with named owners does more for a bill than any dashboard.
What AWS actually gives you, tool by tool
The usual advice is "start with Cost Explorer", which is true and unhelpful, because it does not tell you what the rest of the native stack is for or where each piece stops being useful. Here is the honest version, based on running these tools rather than reading their landing pages.
Read this as a map rather than a checklist. You do not need all of it in week one, and two of these tools solve the same problem at different resolutions.
- ✓ Cost Explorer: the interactive view. Filter and group spend by service, account, tag or usage type, and get Savings Plans and Reserved Instance recommendations. Best for answering a question in two minutes.
- ✓ Cost and Usage Report, now delivered through Data Exports: the queryable truth. Hourly, resource-level, every line item. This is the only source that can answer questions the console cannot express.
- ✓ AWS Budgets: threshold alerting on cost, usage, Savings Plans utilisation or coverage. The point is the alert reaching an owner, not the number itself.
- ✓ Cost Anomaly Detection: machine-learning alerting on unusual spend. Free, and the single highest value-per-minute thing to switch on.
- ✓ Cost Categories: rules that group spend into your business structure, product line, team, environment, without needing perfect tagging everywhere.
- ✓ Compute Optimizer: right-sizing recommendations for EC2, EBS, Lambda, ECS on Fargate and RDS, derived from CloudWatch metrics.
- ✓ Trusted Advisor: a standing checklist including idle and underutilised resources.
- ✓ AWS Organizations plus service control policies: the governance layer that stops the same waste reappearing in a new account.
The three that quietly cost money
The native stack is described as free often enough that the exceptions surprise people, usually on a bill rather than in a planning document. There are three worth knowing before you build anything on top of them.
None of these are expensive in absolute terms. They are worth knowing because they turn a "free" architecture into a small recurring line item you did not forecast, and because the third one changes what advice you can actually act on.
- ✓ The Cost Explorer API is charged per request, unlike the Cost Explorer console which is free. This matters the moment you script a daily cost report, because a naive loop over accounts and days can make thousands of paid calls. Batch your queries and cache the results.
- ✓ AWS Budgets gives you a small number of budgets per account at no charge and then bills per budget per day beyond that, with budget actions priced separately again. Check the current rates on the cost management pricing page before you roll out a per-team budget policy across dozens of accounts, because that is where a rounding error becomes a line item.
- ✓ The Trusted Advisor cost optimization checks are gated behind a support plan. Per the AWS documentation, the full check set is "available to customers with an AWS Business Support+, AWS Enterprise Support, or AWS Unified Operations plan", while Basic Support gets only the Service Limits category plus selected Security and Fault tolerance checks. Cost optimization is not in that free set, so any guide that says to "just look at Trusted Advisor" is assuming a support tier you may not have.
- ✓ CUR itself is free to generate, but you pay for the S3 storage it lands in and for every Athena query against it. Athena bills per terabyte scanned, so partition the data by date and select explicit columns rather than SELECT *.
Cost Explorer or CUR? You want both, for different jobs
This is the most common point of confusion, and picking only one is what makes a FinOps practice feel blind. They are not competing tools, they are different resolutions of the same data.
Cost Explorer is for the question you have right now. Why did last month jump? Which account grew fastest? What would a Savings Plan cover? You get an answer in a couple of minutes without writing anything.
CUR is for the question the console cannot express. Which specific resource IDs drove the NAT Gateway processing charge, hour by hour, grouped by the tag that only half the team applies? That query has no console equivalent, and it is exactly the kind of question that finds real money.
- ✓ Use Cost Explorer for exploration, trend spotting and commitment recommendations.
- ✓ Use CUR through Athena for resource-level attribution, chargeback and anything you need to repeat on a schedule.
- ✓ Group CUR by usage type, not by service. Data transfer and NAT Gateway processing hide inside a bucket called EC2-Other, and a service-level view shows them as one unexplained number.
- ✓ Turn on hourly and resource-level granularity if you can justify the extra data volume. Daily granularity hides the spiky workloads that are usually the interesting ones.
Where the native stack genuinely runs out
Most content on this topic is written by companies selling the alternative, so it is worth being specific about where AWS native tooling actually stops rather than implying it stops everywhere.
For a single-account or small multi-account estate doing its first optimization pass, the native stack is not a compromise. It is sufficient, and you should not spend money on tooling before you have spent the free effort. The gaps appear later, and they are structural rather than cosmetic.
- ✓ Chargeback and showback across many teams. AWS gives you the data, but not the workflow: no per-team portals, no budget negotiation, no invoice-shaped output per business unit. Building that yourself is a real project.
- ✓ Continuous automated action. Compute Optimizer recommends, it does not act. Nothing native right-sizes on a schedule, reclaims idle resources automatically, or manages commitment purchases continuously.
- ✓ More than one cloud, or SaaS spend alongside cloud. The moment Snowflake, Databricks or a large AI provider bill sits next to AWS, native tooling has nothing to say about the total.
- ✓ Engineer-facing cost feedback in the workflow. Cost estimates in a pull request, or spend visible in the tools engineers already use, has no native equivalent.
A 30-day rollout, in the order that works
Sequence matters more than completeness here. Each step makes the next one smaller, and the free steps come first so you can show a real reduction before asking anyone to approve a commitment.
If you only get through week one and week two, you will still have removed real spend and you will have the data to do the rest whenever you get to it.
- ✓ Week 1, data. Enable Cost Explorer and a daily CUR to S3. Switch on Cost Anomaly Detection with alerts routed to a channel a human reads. Activate owner and environment as cost allocation tags, remembering they are not retroactive.
- ✓ Week 2, free wins. Query CUR by usage type. Delete unattached EBS volumes, orphaned snapshots, unassociated Elastic IPs, idle load balancers, and log groups with no retention policy. None of this touches a running workload.
- ✓ Week 3, right-size. Work the Compute Optimizer list against a full business cycle of metrics, not a week. Migrate gp2 volumes to gp3, apply S3 lifecycle policies, and set explicit CloudWatch log retention.
- ✓ Week 4, commit and govern. Only now buy Savings Plans, against the floor of the usage that survived weeks two and three. Set budgets per team, define Cost Categories to match how the business is organised, and put a monthly review in the calendar with named owners.
The pitfalls that actually cost people money
Four failure modes come up repeatedly, and none of them are about picking the wrong tool.
The pattern is the same in each case: the tooling worked, and the operating model around it did not.
- ✓ Committing before cleaning up. Buying Savings Plans against an unoptimized fleet locks in the oversizing for one to three years. It is the single most expensive sequencing mistake available.
- ✓ Postponing tagging. Cost allocation tags only apply from the moment you activate them, so every month you wait is a month of history you can never attribute.
- ✓ Alerts with no owner. An anomaly notification into an unread mailbox is not detection. Route them to the team that owns the spend, not only to finance.
- ✓ Treating it as a project. A one-time cut is a photograph. Without a monthly cadence, costs return to trend within a quarter or two, and you get to do the whole exercise again next year.
Frequently Asked Questions
Can you do FinOps on AWS without buying a third-party tool?
Yes, and for a first optimization pass you should. Cost Explorer, Cost and Usage Reports queried through Athena, AWS Budgets, Cost Anomaly Detection, Cost Categories and Compute Optimizer cover visibility, alerting, attribution and right-sizing. Third-party tooling earns its cost later, typically when you need chargeback workflows across many teams, continuous automated action, or reporting that spans more than one cloud.
Are AWS cost management tools free?
Mostly, with three exceptions. The Cost Explorer console is free but its API is charged per request, which matters if you script daily reports. AWS Budgets gives you a small free allowance per account and then bills per budget per day. The Trusted Advisor cost optimization checks are gated behind a paid support plan. Cost and Usage Reports are free to generate, but you pay for the S3 storage they land in and for every Athena query against them. Check the AWS cost management pricing page for current rates, since the free allowances change.
What is the difference between Cost Explorer and the Cost and Usage Report?
Cost Explorer is an interactive console view for answering a question in a couple of minutes. The Cost and Usage Report is the raw hourly, resource-level dataset delivered to S3 and queried with Athena, for questions the console cannot express, such as resource-level attribution or repeatable chargeback reporting. They are different resolutions of the same data, not alternatives, and a serious practice uses both.
Why does my AWS bill show a large EC2-Other line?
EC2-Other is a usage grouping that typically holds EBS volumes and snapshots, Elastic IPs, NAT Gateway hourly and data processing charges, and data transfer. A service-level view shows it as one large unexplained number. Group your Cost and Usage Report by usage type rather than by service to break it apart, which is the single most useful CUR query most teams have never run.
Do I need a paid AWS Support plan to manage costs properly?
No, but it changes which advice applies to you. The full Trusted Advisor check set, including cost optimization, is documented as available on Business Support+, Enterprise Support or AWS Unified Operations. Basic Support gets only the Service Limits category plus selected Security and Fault tolerance checks, so any guide leaning on Trusted Advisor is assuming a tier you may not have. Note AWS has announced that Developer Support, Business Support and Enterprise On-Ramp are discontinued from 1 January 2027, so verify your plan name against the current support documentation. Everything else in the native stack, including Compute Optimizer and Cost Anomaly Detection, works regardless of support plan.
How long before AWS native FinOps tools show results?
Waste removal produces a visible reduction on the next invoice, usually within the first two weeks, because deleting idle resources takes effect immediately. Right-sizing follows over the next couple of weeks once you have validated it against real metrics. Commitment savings begin the day you buy and compound over the term. The governance work produces nothing measurable in month one and is the reason the savings still exist in month twelve.
Sources
About the author
Hermann Lotter
FinOps practitioner who has led cloud and AI cost optimization inside a 180-person organisation, identifying six-figure annual savings across AWS and LLM spend. He writes Easy Entropy from hands-on engagements, not theory. LinkedIn
Free Assessment
Want this outcome in your AWS bill?
Get a free cloud cost analysis and a prioritized optimization roadmap.
Request Free Analysis →Related Articles
FinOps
How Much Does a FinOps Consultant Cost in 2026? Real Numbers
Published market rates, the three pricing models, and an honest rule of thumb for when hiring cost help pays for itself, and when it does not.
Cost Optimization
7 AWS Cost Cuts You Can Make This Week (2026 Playbook)
A ranked, implementation-first checklist for reducing AWS spend in 2026 by focusing on the fastest, highest-impact savings opportunities first.