A live readout, rendered at the moment you loaded this page. The same figures
as JSON are at /healthz, which is what an uptime monitor
should poll.
Answering
Requests since the last deploy, counted by crawler family from the user agent alone. No paths, no addresses, nothing stored.
Every request to a paid route, by the stage it reached. "Issued" is a 402 answered to a request without payment; "refused" is a request rejected before payment; "attempts" carried a payment header; "settled" is money that landed.
| route | 402 issued | refused | paid attempts | settled |
|---|---|---|---|---|
| POST /v1/domain | 3740 | 0 | 19 | 1 |
| POST /v1/translate | 2429 | 0 | 0 | 0 |
| POST /v1/convert | 2938 | 0 | 1 | 1 |
| POST /v1/render | 1014 | 0 | 0 | 0 |
| GET /v1/convert | 1386 | 0 | 0 | 0 |
| GET /v1/domain | 1903 | 0 | 0 | 0 |
| GET /v1/render | 1588 | 0 | 0 | 0 |
| GET /v1/translate | 1341 | 0 | 0 | 0 |
| route | refused with | count |
|---|---|---|
| POST /v1/domain | 402 payment_rejected | 18 |
Scribe holds no queue and no customer state, so there is no lag to report: a request is answered in its own response or refused before payment. The funnel is the whole story of the service.