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.
Incidents and alerts
Provider failures become incident records, and a webhook tells you about them.
Failures that say something about a credential become incidents. An incident carries the reason, the route and provider, when it started, how long it lasted, and how many attempts were involved. It stays open until the credential recovers or an operator closes it.
What creates an incident
| Signal | Effect |
|---|---|
401 from the upstream |
The credential is retired and stays held until an operator restores it |
| Rate limit / quota exhausted | The credential cools down for the configured window, then rejoins selection |
| Repeated upstream unavailable | The credential cools down and keeps being probed |
| Recovery | The incident closes when a probe succeeds |
Errors caused by the request itself (for example an upstream 400) do not create
incidents and never remove a credential from rotation.
Alerts
The only delivery channel today is a webhook. Configure it under System → Alerts:
| Setting | Meaning |
|---|---|
| Webhook URL | Where incident events are POSTed; empty disables alerts |
| Notify on recovery | Also send a payload when an incident closes |
| Minimum interval | Per-provider throttle so one flapping account cannot flood the channel |
Delivery is asynchronous and never blocks forwarding: 5-second timeout, up to three retries with backoff, and no credentials in the payload.
A webhook is the only channel today; point it at a relay if you need another transport.
In the console
Provider incidents lists open and closed incidents with filters for route, provider, severity and time range. Retired credentials can be restored from the route page, which also closes the incident.