Where publishers disagree
Providers for which independent publishers report a materially different error-code profile (a different dominant code, or a >2× spread in total errors). This is the most interesting output of a federated health commons — no single source of truth, and every row is independently signature-verified so you can trust who said what. Codes are provider-specific (429 rate_limit, 529 overloaded, 503, insufficient_quota, …) — look them up in that provider's docs.
No disagreements right now — all publishers agree on every provider they report, or the index is still warming up.