Solidway v0.1.0 is released under AGPL-3.0; source and releases are on GitHub. This documentation is being rewritten against the release — pages outside Quick start and Reference still describe planned behaviour. The current reference lives in docs/getting-started.md and docs/architecture.md in the repository.
Usage and metrics
What the console counts, at what granularity, and how long it is kept.
Solidway counts usage itself and shows it in the console. There is no external metrics stack to run, and no scraping endpoint to configure.
What is counted
| Metric | Broken down by |
|---|---|
| Requests, completed, errors | route, API key, provider, model, owner |
| Tokens — input, output, cached, reasoning | same dimensions |
| Latency — average, p95, time to first byte | same dimensions |
| Failovers | route and attempt |
| Credential health | per credential, live |
The console has two scopes: Usage for the routes you own, and Global usage for admins across all users.
Granularity and retention
Two layers, both configured under System → Usage:
- Aggregates — per-hour buckets, always on, cheap to keep for a long time.
- Per-request detail — a row per request, with a mode: off, errors and failovers only, or all requests. Retention hours and a sample rate keep the detail database bounded.
Time windows
Presets run from the last hour to the last 30 days, plus custom ranges. Buckets align to the server’s timezone offset, so “today” means your today, not UTC’s.
What is not here yet
There is no Prometheus endpoint, no OpenTelemetry/OTLP exporter and no metrics push. If you need them, the per-request detail database is the place to attach a collector.