files.brainforge.ai outage investigation (2026-03-18)

Summary

  • Reported issue: https://files.brainforge.ai/sales/ai/brainforge_ai_capabilities_deck.pdf was inaccessible.
  • Investigation result: this is a domain-level outage for files.brainforge.ai, not a single missing file.
  • Most likely root cause: files.brainforge.ai is pointed at a Heroku app endpoint that currently has no active app/domain binding.

Trigger / context

  • Slack thread in #brand-marketing-design reported broken capability deck link and concern that all links might be down.

Checks performed

1) DNS resolution

  • dig files.brainforge.ai CNAME +noall +answer
  • Result:
    • files.brainforge.ai -> metric-bat-nht5fk8ijris9q3smo7olvoe.herokudns.com
  • Interpretation: domain currently targets Heroku DNS, not Railway edge.

2) HTTPS behavior

  • curl -I https://files.brainforge.ai/...
  • openssl s_client -connect files.brainforge.ai:443 -servername files.brainforge.ai
  • Result:
    • TLS handshake fails with tlsv1 unrecognized name / no peer certificate.
  • Interpretation: HTTPS for this host is not correctly provisioned for the current target.

3) HTTP behavior

  • curl -I http://files.brainforge.ai/
  • curl http://files.brainforge.ai/
  • Result:
    • HTTP/1.1 404 Not Found
    • Body title: No such app (Heroku error page)
  • Interpretation: Heroku router is reachable, but no corresponding app/domain mapping exists.
  • Sampled multiple https://files.brainforge.ai/...pdf links found in repo references.
  • Result:
    • HTTPS calls failed similarly.
    • HTTP fallback returned 404 for sampled asset paths.
  • Interpretation: impact is broad across files.brainforge.ai links.

5) Railway status/log attempt

  • Railway CLI installed in agent environment.
  • Authentication could not be established with available token in this runtime; direct Railway deployment/build logs could not be pulled from this session.
  • Additional signal:
    • labs.brainforge.ai resolves to *.up.railway.app and responds via railway-edge, indicating Railway platform itself is reachable for other services.

Conclusion

  • The outage is most consistent with custom domain misconfiguration or decommissioned app on Heroku for files.brainforge.ai.
  • This does not look like a single bad PDF path.
  1. Decide canonical host for files.brainforge.ai (Heroku vs Railway).
  2. If Railway is canonical:
    • Attach files.brainforge.ai custom domain to the intended Railway service.
    • Update DNS CNAME from Heroku target to Railway-provided target.
    • Verify TLS certificate issuance and 200/302 response for known files.
  3. If Heroku is canonical:
    • Restore/recreate app serving assets.
    • Reattach custom domain and ensure ACM certificate is active.
  4. Post-fix validation:
    • Test 5-10 representative URLs (sales deck, case studies, comparisons).
    • Confirm HTTPS works (no TLS/SNI errors) and links return expected content.

Suggested prevention

  • Add scheduled synthetic checks for a small set of critical files.brainforge.ai URLs with alerting on:
    • TLS handshake failure
    • 4xx/5xx responses
    • unexpected redirect chains