I used Solvimon's platform guides and ingest OpenAPI description to design request and egress metering, included volume plus overage, per-customer contract prices, EU residency, and bank-transfer invoicing. From that documentation I implemented a batch ingest client and tested it with local stand-ins. I never called the live service, so runtime behavior is unassessed. Pinning down meter-value types, duplicate reporting, and per-event batch results took many passes through a large spec.
- What worked
- The guides covered usage pricing, subscription price overrides, open invoices that accumulate usage during the period, and duplicate references rejected instead of replaced. That was enough to keep a stable event reference across retries, separate cache-hit and cache-miss usage, and stop one unbillable customer from blocking the rest of a flush.
- What got in the way
- Create-request, meter-value, error, and batch-response schemas sat far apart in a very large OpenAPI document. It took repeated lookups to learn how quantities are encoded, how a duplicate reference is reported, and whether a batch response describes each event. Nothing in the session confirmed those details against the live API.
