Custom domain www.logic-motive.com (service: web, project: insightful-reverence,

environment: production) has a confirmed backend routing bug:

• Domain binding is silently stripped from service config after every deployment

• When present, routing points to STALE/dead deployment instances instead of the

  current live one (confirmed instance IDs: dead=4e4c9caa-9c02-420d-bd29-617aa5db5ea9,

  live=49b3e79e-b638-4680-909a-fe58c236dd09)

• Requests to www.logic-motive.com return 502 directly from Railway edge

  (x-railway-fallback:true, server:railway-hikari) — request never reaches container

• TLS handshake completes cleanly (curl -v confirms), ruling out cert issues

• Default domain (web-production-28e4f.up.railway.app) works perfectly (200 OK on /healthz/)

• DNS independently verified correct on 8.8.8.8 and 1.1.1.1:

  www.logic-motive.com CNAME -> n3x383ud.up.railway.app

• Reproduced across 3+ delete/re-add cycles over 3 days

Please manually clear/reset the routing cache for this custom domain and re-bind it

to the current live deployment instance.