8 min read · Updated 2 Nov 2026
Why did my AWS bill go up? Find the cause in 30 minutes
In short: A step-by-step way to find what made your AWS bill jump: the day it started, the service and usage type behind it, the resource, and the change or deploy that caused it.
Free tool: Kubernetes waste calculator → · Log cost calculator →
An unexpected AWS bill usually has one or two causes, not twenty. The trick is to narrow down in the right order: when did it start, which service, which usage type, which resource, and what changed just before. Here's how to do that with the tools AWS already gives you.
1. Find the day it started
In AWS Cost Explorer, set the time range to the last two or three months, granularity to Daily, and group by Service. A jump that starts on a specific day points to a change; a slow climb points to growth, such as more data, more traffic or more logs. Note the date. Everything after this is about that date.
2. Find the service and the usage type
Click into the service that grew and group by Usage type. The usage type tells you what you're paying for, which is much more useful than the service name. A few that often explain a surprise:
- NatGateway-Bytes: data processed by a NAT gateway, often traffic to S3 or other AWS services that could use a free gateway endpoint instead.
- DataTransfer-Out-Bytes: data leaving AWS to the internet.
- BoxUsage: EC2 instance hours, such as new instances, an autoscaling group that grew, or instances that were never stopped.
- EBS:SnapshotUsage: snapshot storage that grows silently.
- TimedStorage-ByteHrs under CloudWatch: log storage, which never expires unless you set retention.
- PublicIPv4: AWS charges for every public IPv4 address, including ones attached to nothing.
3. Find the account, region and resource
Group by Linked account and by Region to see where the cost is. To get to the individual resource, you need the Cost and Usage Report (CUR) with resource IDs turned on: it lists every line item with the resource ID, so you can sort by cost and see the exact instance, volume, bucket or NAT gateway.
4. Find what changed
- CloudTrail event history shows the last 90 days of management events: who created, modified or deleted which resource, and when. Search for create and run events around the day the cost started.
- Check your deploy log and Terraform history for the same day: a new service, a changed instance type, a scaled-up cluster, a debug log level left on.
- Check whether a Savings Plan or Reserved Instance expired, which turns the same usage back to on-demand prices overnight, or whether a free tier period ended.
5. Fix it, then stop the next one
- Fix the cause, not the symptom: a VPC endpoint instead of NAT traffic, log retention instead of deleting logs, a smaller instance instead of a bigger budget.
- Turn on AWS Cost Anomaly Detection so you get an alert the day spend jumps, not at the end of the month.
- Tag resources by team and environment, and activate the tags for cost allocation, so the next jump has an owner straight away.
CostLeak automates steps 1 to 4: it reads your Cost and Usage Report, finds the day spend jumped, splits the increase by service, usage type, resource and team, and lines each part up with your deploys and CloudTrail events.
FAQ
Questions
How quickly does Cost Explorer show new costs?
Usually within about a day. The Cost and Usage Report is updated up to several times a day and finalised after the month closes.
Why is my NAT Gateway so expensive?
NAT gateways charge per GB processed. Traffic from private subnets to S3 or DynamoDB can use free gateway endpoints instead, which often removes most of the charge.
How far back can I see who made a change?
CloudTrail event history keeps 90 days of management events per region. A trail that delivers to S3 keeps them for as long as you retain the logs.
Can I get alerts before the bill arrives?
Yes. AWS Budgets alerts on thresholds, and Cost Anomaly Detection alerts on unusual daily spend by service or account.
Tools for this
Products that help
CostLeak
AWS cost leak finderFind the money AWS keeps. Reads your AWS bill and finds money being wasted, such as unused servers, forgotten storage and costly network traffic, with the yearly saving for each and what to do.
- Who it's forStartups and SMBs spending $5k–100k a month · Agencies hosting client accounts
- What you getSavings report with yearly amounts · Savings Plan sizing · Prioritised action plan
DriftTrail
Drift detectionKnow who changed production. Finds changes someone made by hand in AWS, Azure or Google Cloud that aren't in your Terraform code, tells you who made each one and when, how risky it is, and whether to undo it or keep it.
- Who it's forPlatform teams where engineers still have console access · Security teams preparing for SOC 2, ISO 27001 or RBI audits
- What you getEvery drift with a name and time · Revert-or-adopt advice per change · Nightly drift check in CI
PodRight
Kubernetes right-sizingPay for the Kubernetes you use. Looks at how much CPU and memory your Kubernetes apps really use, then tells you how much to cut (and save each month) and which apps are about to crash from too little.
- Who it's forCTOs and platform leads on EKS, GKE or AKS · FinOps owners
- What you getRight-sizing report with savings · Risk list (OOM, throttling, eviction) · kubectl commands and patches