Official status or synthetic probes: which signal do you trust first?

I'm building OutageDeck around official provider status feeds, but I do not think an official feed should be the only production signal. Providers can acknowledge incidents late, while synthetic checks can fail because of your own credentials, DNS, region, or deployment.

My current triage model is:

- provider incident + probe failure: strong upstream evidence

- provider operational + probe failure: unclassified failure, investigate our stack and keep watching the feed

- provider incident + probe passing: warn, but do not page the same way

I made a no-account stack view for the provider side: https://outagedeck.com/stack?utm_source=producthunt&utm_medium=forum&utm_campaign=prelaunch_two_signal_discussion

For people who have handled real incidents: which signal has misled you more often, a green provider dashboard or a synthetic probe that blamed the wrong dependency? What evidence do you want attached to the first alert?