Custom Domains Not Working from Russia

I am experiencing a critical issue with my custom domains that has persisted for several days, and I have exhausted all possible troubleshooting steps on my end.

Affected Projects & Domains:

  • Project: rtk-client-nojm | Custom Domain: telecom-rtk.ru | Vercel URL: rtk-client-nojm.vercel.app (works perfectly)
  • Project: mts-r | Custom Domain: mts-s.ru | Vercel URL: mts-r.vercel.app (works perfectly)

Core Problem:
Custom domains either fail to load entirely or load intermittently with ERR_CONNECTION_RESET 200 (OK) errors, while all corresponding .vercel.app domains work flawlessly.

Detailed Investigation Performed:

  1. DNS Configuration:

    • Verified DNS records are correct across multiple providers (Timeweb)
    • Tried multiple A records (76.76.21.21, 76.76.21.22, 216.198.79.1, 216.198.79.2)
    • Confirmed global DNS propagation via DNS checker tools
  2. Network Analysis:

    • Requests reach Vercel servers (Frankfurt region based on x-vercel-id: fra1)
    • Server responds with 200 OK and begins sending data
    • Connection resets mid-transfer with TCP RST packets
    • Issue occurs across different Russian ISPs (Rostelecom, MTS)
  3. Architecture Changes:

    • Recently moved some backend APIs from Render to Vercel
    • Frontend makes requests to Vercel-based APIs
    • Confirmed no CORS issues (requests work via .vercel.app domains)
  4. Geographic Testing:

    • Sites work reliably from other countries and most Russian cities
    • Intermittent success - sometimes loads after multiple refresh attempts
  5. Headers Analysis:

    • x-vercel-cache: MISS consistently
    • cache-control: private, no-cache, no-store headers present
    • Successful TLS handshake, connection established then reset

Critical Observations:

  • Problem started around October 20-21, 2024
  • .vercel.app domains remain 100% functional
  • Custom domains show inconsistent behavior
  • No application-level errors, pure network-level connection resets

Request:
Please investigate:

  1. Possible regional routing issues to edge nodes serving custom domains
  2. Any firewall/rate limiting differences between custom domains and .vercel.app domains
  3. TCP connection timeout settings for Russian IP ranges
  4. Recent infrastructure changes that might affect custom domain routing

The fact that .vercel.app domains work perfectly while custom domains fail suggests a infrastructure-level issue specific to custom domain handling.

Thank you for your urgent attention to this critical matter.

Primary Contact: telecom-rtk.ru
ISPs Affected: Rostelecom, MTS

The domain troubleshooting guide can help with most custom domain configuration issues. You might be able to use that guide to solve it before a human is available to help you. Then you can come back here and share the answer for bonus points.

You can also use v0 to narrow down the possibilities.

Custom domains telecom-rtk.ru & mts-r.ru failing with ERR_CONNECTION_RESET from Russia, while .vercel.app domains work perfectly.

Issue: Custom domains load intermittently or fail completely from Russian ISPs (Rostelecom, MTS). Server responds with 200 OK but connection resets mid-transfer.

KEY FINDING: Sites work instantly via VPN or from other countries. Problem is specific to Russian IP ranges.

What I’ve verified:
:white_check_mark: DNS correct (multiple A records: 76.76.21.x, 216.198.79.x)
:white_check_mark: Global DNS propagation confirmed
:white_check_mark: Reaches Vercel servers (Frankfurt - fra1)
:white_check_mark: Works via VPN/outside Russia
:white_check_mark: No CORS issues (.vercel.app domains work)

Critical observation: Problem started Oct 20-21. .vercel.app domains & VPN work 100% while custom domains from Russia fail.

Request: Investigating routing/firewall issues to edge nodes for custom domains from Russian IPs, or recent infrastructure changes affecting this region.

Any help appreciated!

I recommend going through this:

Everything has already started working on its own, I don’t know what will happen next, but it seems that the problem was on the side of one of the services. I don’t know how things will work out, but I hope the problem doesn’t happen again.

Great to hear it’s all good now! Let us know if you have any more issues.

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.