IOTA Watch: validating parallel upstream arrays before rendering rewards

[molimao]

I maintain IOTA Watch, an MIT-licensed training and mining monitor. One small Zod pattern in the project may be useful to other dashboards consuming external APIs.

Some upstream endpoints return several parallel arrays: an amount, timestamp and status at the same index form one reward record. Validating each array on its own is insufficient, because unequal lengths can silently pair the wrong values. Our helper builds a z.object(shape) and adds a .refine() requiring all listed array lengths to match.

The schemas also distinguish required numeric fields from optional or nullable metadata. Counts must be finite and nonnegative; loss is allowed to be null. A malformed payload produces an unavailable result instead of a plausible-looking zero. Successful previous data can be retained with its original source clock while the failure remains visible.

Schema definitions and alignment helper · data-reliability tests

This is a self-submitted application example, not a Zod extension. The UI is available in five languages and keeps reward units and accounting periods separate.