We are seeing intermittent outbound TCP connection failures from a Railway service in the US West region to an Alibaba Cloud Model Studio endpoint in Beijing.

Environment: production

Region: US West

Destination host:

ws-douf5b4ymvzf09du.cn-beijing.maas.aliyuncs.com

Port:

443

DNS resolution succeeds and returns both IPv4 and IPv6 records.

From the same deployed Railway container:

• https://example.com consistently returns HTTP 200
• requests to the Alibaba endpoint intermittently succeed and return HTTP 401 when unauthenticated
• other attempts fail during TCP connect with IPv4 ETIMEDOUT
• IPv6 connection attempts report ENETUNREACH

Example results:

• HTTP 401 in 1241 ms
• TCP connect failure in ~505 ms:
  • IPv4: ETIMEDOUT
  • IPv6: ENETUNREACH
• HTTP 401 in 223 ms

This appears to be an intermittent outbound connectivity/routing issue rather than a DNS or application configuration issue, since the same endpoint sometimes returns a valid HTTP response.

Could you please help investigate whether there are any outbound routing, IPv4, or IPv6 connectivity issues from Railway US West to this destination?