Alerts

Alerts let you watch any Spanly metric – error rate, latency, request volume, tool-call failures – and get notified when it crosses a threshold you care about. Rules are scoped to a project and can fire on a single server, a single tool, a client, or the project as a whole.

Notification channels

You can deliver alerts to:

• Email – any user in the project's team.
• Slack – install the Spanly Slack app and pick a channel per rule.
• Webhook – POST a JSON payload to any HTTPS endpoint for routing into PagerDuty, Opsgenie, or your own incident system.

Typical rules to set up

• Tool call error rate above 5% over 5 minutes on a critical server.
• p95 request duration above 2s on the tools/call method.
• No traffic from a production client for 10 minutes (deadman switch).
• A specific tool – e.g. delete_file – invoked more than N times in an hour.

Status

Alerts is rolling out – the API surface and dashboard configuration are still evolving. Check the Alerts tab in your project for the current set of supported conditions, or email support@spanly.com if you need a rule type that isn't there yet.