noy-db
stable

probe

Store diagnostics, shipped in @noy-db/to-meter since the @noy-db/to-probe package was retired — benchmarks any NoydbStore and scores its role suitability.

Ships in @noy-db/to-meter · pnpm add @noy-db/to-meter Not a backend — a 5-axis benchmark you point at any store.

@noy-db/to-probe is retired. The package is deprecated on npm and receives no releases. Its two entry points — runStoreProbe() and probeTopology() — were absorbed into @noy-db/to-meter in 0.4.0-pre.8; import them from there. A diagnostic that returns no store never fitted the to<Backend>() store contract, and probing to choose a store then metering to watch it are one workflow.

import { runStoreProbe, probeTopology } from '@noy-db/to-meter'
Ships in@noy-db/to-meter
Shapen/a — diagnostic, returns no store
Wrapsany NoydbStore

What it does

runStoreProbe exercises the 6-method NoydbStore contract across five axes — write latency, CAS integrity, hydration cost, sync economics, and network resilience — and produces a structured, risk-scored suitability report for the roles primary, sync-peer, backup, and archive. probeTopology runs a setup-time topology check.

Notes

  • Each suitability.risks entry carries a machine-readable code ('slow-write-p99', 'cas-unsupported', …), a severity ('warn' or 'error'), and a human-readable message.
  • Any error-severity risk empties recommended entirely; slow-write-p99 drops primary and cas-unsupported drops sync-peer — other warn-level findings are reported without affecting the recommendation.
  • Doesn't persist anything itself — it's diagnostic tooling, not a storage tier.

See also

postgres

rest