What the page speed watcher costs
This is the most compute-heavy system in the series, because every measurement is a real browser loading a real page. Eight pages measured nine times a day is eighteen hundred page loads a month, each a few seconds of Lambda. Here is where each cent goes.
Key takeaways
- About $12 a month at 24 pages daily. Roughly $33 at 72 pages daily.
- One Bedrock read per page load is the only line that scales. Everything else is rounding.
- Nothing is always-on, so a quiet month genuinely costs almost nothing.
- Browser time is the whole bill. Nine samples is nine times the cost of one, and it is what makes the numbers usable.
- 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 pages daily | How it scales |
|---|---|---|
| Bedrock read | $0.00 | Linear. One call per page load, roughly 1,800 in and 200 out tokens. |
| Lambda | $2.27 | Linear at $0.0004 per page load. |
| SES | $2.71 | Linear. About 0.01 messages per page load. |
| DynamoDB + S3 | $1.14 | 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 compute band is Lambda duration running a headless browser, at roughly two to four seconds per page load on a throttled profile. There is no model in this system, so there is no read line at all.
The three ways this bill surprises you
Every one of these has happened to somebody, and all three are cheap to prevent.
- Measuring too many pages. The cost is linear in pages times samples, and a list of eighty pages measured nine times daily is a genuinely noticeable bill for a report nobody can act on. Eight pages is not a limitation; it is the design.
- A browser that fails to close. A headless browser left running until the Lambda times out costs the full timeout every invocation. Always close in a finally block, and set the timeout to roughly twice the expected run.
- Storing full HAR files forever. A complete request trace per sample is megabytes. Store the resource summary always and the full trace only for runs that regressed.
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