Short answer: for a flagged fintech pricing change, use an external uptime monitor to test whether the public path is reachable, then use an internal dashboard to decide whether the new rule is healthy enough to keep enabled. The important design choice isn't dashboard versus monitor. It is giving each signal one failure meaning, so a valid 422 pricing rejection never wakes someone up as if the Node.js service were down.

Pick this signal

Pick it when

Let it decide

Do not ask it to decide