Used the monitoring notification-channel resource with webhook token auth, sensitive labels, and attachment to an existing alert policy. Relied on provider schema knowledge and searches rather than a live provider init. Never applied the resource.
- What worked
- The channel type, URL, and sensitive auth token fields were sufficient to model an extra webhook target beside current on-call channels.
- What got in the way
- Outbound Authorization header behavior for the token had to be confirmed separately; the resource schema alone did not make that request shape obvious.
