RuntimeError

What happened

• The nightly billing job (about 3am) submits encounters to Candid. Candid answers HTTP 422: the provider needs an organization_name or a first and last name.
• The mapper builds billing_provider from the Candid provider record for the program's claims_billing_provider_uuid. It uses organization_name and drops nils with .compact, so a record without a name sends an empty name.
• handle_submission_error calls Bugsnag.notify on purpose, so this is a deliberate warning and not a crash.
• There are 1181 events over 43 days with 0 users. That fits the same few entries failing and being retried every night, not a deploy blip.

What to do next

1. Query Lists the billing provider UUIDs configured per program so you can check each one in Candid for a missing organization_name.

   SELECT program_id, claims_billing_provider_uuid FROM program_settings WHERE claims_billing_provider_uuid IS NOT NULL;

2. Check In the Candid dashboard, open the organization provider for each UUID from the query above. Make sure the failing one has an organization_name, or first and last name, and fix it there. Also confirm the rendering provider NPI record is complete.
3. Code fix In app/lib/candid_health/care_management_claim_mapper.rb, add a validation step for the billing provider. If the record has no organization_name and no first_name plus last_name, return a mapper validation error (the path that calls record_validation_failure) with a clear message. Do this before submitting to Candid.

Queries are shown, never run. Check them before using them on production.