What the budget variance reporter costs
This system runs once a period over a few thousand rows, which makes it the smallest workload in the series by some distance. Twelve periods a year for one entity is twelve runs; a group with five entities is sixty. Here is where each cent goes.
Key takeaways
- About $2 a month at 24 periods. Roughly $3 at 60 periods.
- One Bedrock read per period is the only line that scales. Everything else is rounding.
- Nothing is always-on, so a quiet month genuinely costs almost nothing.
- A period is a few thousand rows and the system runs monthly, so the annual bill is smaller than most single subscriptions it might find.
- The duplicate test runs before the read, so resends are free.
- The three real risks: a retry loop, storage nobody expires, and a bigger model than the job needs.
The bill at three volumes
These are US East prices at the time of writing, at three volumes that bracket most small businesses. Find the bar closest to your own and read across.
Line by line
| Line | At 24 periods | How it scales |
|---|---|---|
| Bedrock read | $0.50 | Linear. One call per period, roughly 1,800 in and 200 out tokens. |
| SES | $0.01 | Linear. About 1 messages per period. |
| DynamoDB + S3 | $0.28 | Storage grows with what you retain, not with throughput. |
| Lambda + SQS | $0.12 | Linear, and effectively free at this scale. |
| CloudWatch | $0.16 | Flat, if you set retention. Unbounded if you do not. |
| Secrets Manager | $0.40 | Flat. One secret, $0.40 a month. |
| AWS Budgets | $0.46 | Flat. Two actions, so you find out before the bill does. |
The read cost per period is higher than elsewhere in this series because the prompt carries the surviving variance lines and their transactions. It is still about two cents, and it runs once a month.
The three ways this bill surprises you
Every one of these has happened to somebody, and all three are cheap to prevent.
- Re-running on every restatement. Restatements are normal, but a system that re-imports and re-analyses on every identical re-send pays repeatedly for nothing. The file digest check is what prevents it.
- Sending the whole period to the model. The model is only needed to write the explanation sentences for the three or four surviving lines, not to read thousands of transactions. Feeding it the period is a hundredfold cost increase for no gain.
- Log retention left at never. This system runs twelve times a year and will otherwise be almost entirely a CloudWatch bill within eighteen months.
What it costs when nothing happens
This matters more than the headline number for a seasonal business. In a month with nothing to process the bill is the fixed band: Secrets Manager at forty cents, AWS Budgets at forty-six, and a few cents of storage. Call it a dollar. There is no instance to stop and nothing to remember to turn off.
- Management
- Analytics
- Front-end & mobile
Set these on day one
- A dead-letter queue on every SQS queue, with a maximum receive count of three.
- Thirty-day retention on every CloudWatch log group. There is no default that is safe.
- An S3 lifecycle rule on the object prefix, tiering at 90 days and expiring at your actual record-keeping horizon.
- Two AWS Budgets actions — one that emails at half your expected spend, one at double it. The second is how you find out about a loop in an hour instead of a month.
- Provisioned concurrency: none. Nothing here is latency-sensitive enough to justify paying for a warm function.
Next: the same system drawn for engineers — service names, resource identifiers, IAM scopes, table schemas and the model id.
All posts