Merged into Kong's OpenMeter
Three fixes in the billing engine. A zero-usage tier read from an unsorted price (5347), a prior discount never netted out of a split line (5367), and a package count that rounded past its own boundary (5372).
Billing correctness research
Billed $50. Expected $150.
The issue reporter says the dry run repeated a wrong $50 invoice. Our released-tag H2 test reproduced the $50 invoice. We did not run the dry-run endpoint.
Fourteen findings in five billing engines. Nine landed, one was declined, the rest are still open. No customer, no deployment.
August 1 – September 1, 2026
Unseen$100
The preview and the invoice agreed. The approved price did not.
What happened
$150 for the month.
Dry run and invoice: $50.
An independent check of the approved price, before issuance.
The record
Three fixes in the billing engine. A zero-usage tier read from an unsorted price (5347), a prior discount never netted out of a split line (5367), and a package count that rounded past its own boundary (5372).
Nine lines now in OpenMeter's AGENTS.md, constraining how division and rounding may be used before a boundary decision. Written here, merged there, and binding on everyone who writes in that engine next.
Three billing defects, each closed as completed: a campaign budget type the API accepted but no code enforced (17268), a null quantity satisfying a required validation (17513), and an override price list ignored when a sale list sorted first (17529).
Every line above links to a public pull request or issue in a repository nobody here controls. None of it is a testimonial, a logo, or a customer. It is the work, where it landed. How all three were found, and one that costs nothing and was reported anyway.
The ledger
| Project | Finding | Outcome |
|---|---|---|
| OpenMeter | Zero-usage tier read from an unsorted price #5347 | Merged 26 Oct |
| OpenMeter | Gross amount summed where net was promised #5367 | Merged 27 Oct |
| OpenMeter | Package count rounded past its own boundary #5372, plus a rounding rule added to their contributor guide | Merged 27 Oct |
| OpenMeter | Iterator index and timestamp naming different instants #5521 | Fixed by a maintainer, closed 19 Nov |
| OpenMeter | One-time fee prorated against the item period #5373 | Declined. A maintainer named a reasonable competing semantic |
| OpenMeter | Full flat fee billed for a truncated span #5388 | Open, unreviewed since 28 Oct |
| OpenMeter | Patches treating "not active now" as "scheduled later" #5389 | Open. Another contributor is writing the fix |
| OpenMeter | UNIQUE_COUNT counting a missing property as a value #5390 | Open, unreviewed since 28 Oct |
| Medusa | Campaign budget type accepted by the API, enforced nowhere #17268 | Closed as completed |
| Medusa | Null quantity satisfying a required validation #17513 | Closed as completed |
| Medusa | Override price list ignored when a sale list sorted first #17529 | Closed as completed |
| Kill Bill | Colliding catalog price changes at one boundary #2320 | Open, no maintainer review |
| Kill Bill | Same-day catalog change applied to the wrong cycle #2321 | Open, no maintainer review |
| ScopeBlind | Conformance test vectors, five changes | Merged, credited in the README |
Five open and one declined are on this list on purpose. A record that shows only the wins is an advertisement, and the open rows are the honest part: most findings wait, and some are wrong.
Where this stands
Open-source checks for completeness, window, and truncation, with a published method. The HTTP decision API and Observe Mode wrapper are implemented and tested locally. They are not deployed or running in anyone's production path.
Nobody is blocking live orders through Nobulex. The separate gateway prototype returns permit, block, or escalate and has no customer result. Nothing is published from the registry; every issued record is held.
On the measured corpus a thirty-line in-house version catches the same thirteen faults and blocks none of the clean cases. The only margin is abstention, five of five. Selling the gateway would need an adoption claim the record does not have, so it stays a prior prototype and the audit is the work.
Questions
A proposed study of one billing change. Not a deployed billing product, and not a customer result.
No. The checks are open source. The gateway is local. This page will keep saying so until that changes.
A DMCA notice, and it was correct. One test file in it was a byte-identical copy of an AGPL-licensed file, redistributed under MIT with the author's notice stripped. That was my mistake. The removal is prepared, it covers ten files rather than the one named, and the repository stays down until GitHub acts. The open-source checks and the method are unaffected and still published.
No. A signature shows who signed a decision. It cannot turn a wrong source into a right one.
Billing accuracy for usage-based products. A fixed-fee audit now, a read-only scanner for self-contradictions in design, mutation checks in CI after that, and each layer gets built only once the one before it pays. The approach, with its limits stated.
Next
Tell us how you check invoice amounts before release.
nobulex.dev@gmail.com