Checklist
Five lines to check on your cloud bill this week
The five lines on an AWS bill where I find waste first, with where to look in Cost Explorer and what to change.
I'm starting to see a pattern in startup cloud bills. Containers on a managed runtime and one managed Postgres database carry a lot of teams to Series A. The bill stays small as long as someone reads it. These are the five lines where I find waste first, in the order I check them.
Everything starts with one Cost Explorer export. I take the last three months, monthly, grouped by service and then by usage type. Most of the answers sit in that one file.
From the talk: Lean cloud architecture that scales
The five lines
Each yes scores one point.
- 1.Idle non-production environments. Staging, preview, and dev stacks running all 168 hours of the week. A schedule for a 50-hour working week cuts their cost by about 70%. Some services scale to zero, which cuts it further.
- 2.NAT gateway data processing. The NatGateway-Bytes usage type. AWS charges every gigabyte through a NAT gateway on top of the hourly fee. Gateway endpoints carry S3 and DynamoDB traffic at no charge.
- 3.Oversized databases. The RDS instance class against its CPU and memory use over the last month. Unused provisioned IOPS, and Multi-AZ on staging, where a restore from backup is usually enough.
- 4.Old snapshots and unattached volumes. EBS volumes in the available state, snapshots older than the retention period, and old AMIs with the snapshots behind them. Each one is small. Together they rarely are.
- 5.Unclaimed credits and discounts. Startup credits (such as AWS Activate) tracked against their expiry date. Savings Plans only after two to three months of stable usage, so the commitment matches real usage.
How I export the bill in five minutes
- I open Cost Explorer in the management account. Range: the last three full months. Granularity: monthly.
- I group by service. That CSV shows where the money goes.
- I group by usage type. The second CSV shows why: hours, gigabytes, requests, or storage.
- I sort both by the latest month. Any line that grew faster than the customer count gets a mark.
What I leave out on purpose
I leave out Kubernetes, a service mesh, and multi-region active-active. Each one adds a full-time job a small team doesn't have yet. The bill gets bigger and harder to read with every one of them.
Next step
I highly recommend bringing that Cost Explorer export to a thirty-minute call. I'll point at the lines I'd look at first. Which line grew fastest last quarter?
Book a callThese takeaways are an engineer's reading of the rules and the tools. I wrote them to help teams decide what to build. They aren't legal advice. I recommend checking the current text with your own counsel before planning around a date or a threshold.
Other takeaways from this talk
Checklist
Day-one foundation checklist
Six yes/no checks for a cloud foundation that passes a customer's security questionnaire in an afternoon instead of a quarter.
Scorecard
AI workflow payback scorecard
Four questions that tell whether an AI workflow is worth building before anyone picks a model, scored out of eight.
Let's talk.
A 30-minute call, no pitch. We look at your current setup, your stage, and the thing that's bothering you about it. I give a straight opinion on what to change, what to leave alone, and whether the problem is the one it looks like. What's the one thing in your stack you'd fix tomorrow?
Prefer email or phone? hello@netninjas.eu or +31 6 45984313