How-to
How to see AWS cost per resource, not just per service
AWS Cost Explorer groups spend by service. Here's how to attribute cost to individual resources — this bucket, that instance — so you know what's actually expensive.
AWS Cost Explorer will happily tell you that you spent $340 on EC2 last month. What it won't tell you, at a glance, is which instance. When spend is grouped by service, the one runaway resource hides inside the total, and you end up guessing.
Per-resource attribution is the thing you actually want, and here's how to get it.
Why service-level cost hides the problem
A service total is an average of everything. Three cheap instances and one oversized one all roll into "EC2." The signal you need — this specific resource is the expensive one — is exactly what the grouping erases. To act, you have to break the total back down by resource, which usually means exporting the Cost and Usage Report and writing queries.
The resource-level view in Korve
Korve samples the running state of your connected account and attributes cost to individual resources. On the Cost page and in the resource list, a running EC2 instance or an available RDS database carries its own estimated cost — a per-resource chip, not just a service bucket. You see a 14-day spend trend, this month's total, and which resources are driving it.
Because the resources are the same nodes you see on the canvas, the expensive one isn't an opaque ID in a spreadsheet — it's a node you can open, resize, or delete.
A practical workflow for cutting spend
- Open the Cost page and look at the trend. A step up usually means something was launched and left running.
- Sort the resource list by estimated cost. The offender is near the top.
- Open that resource. Is it oversized? Idle? Forgotten from a test?
- Right-size or terminate it from the same view, then re-check the trend over the next day.
Most surprise AWS bills are one or two forgotten resources, not a broad increase. Per-resource attribution turns "why is this so high" into "oh, that."
A note on billing lag
AWS billing data lags roughly a day, so today's spend isn't final until tomorrow. Korve's estimates for running resources are based on their current configuration and uptime, which is why a just-launched instance shows an estimate immediately while the official number catches up.
Related
Build this on a canvas instead of the console.
Start Korve free →