a month ago
Since today (13 Oct) our outbound traffic from Railway is being treated as high-risk by two providers, and it has taken our product's AI and CRM features offline.
What happened
- 13:05 UTC, Salesforce: our workflow-worker ran its normal scheduled Salesforce API sync. Salesforce classified the source IP 152.55.185.5 as a "high-risk anonymizing proxy (e.g. TOR, Mullvad VPN)", froze the users and revoked their OAuth tokens. The same sync ran fine about 20 minutes earlier.
- 13:31 UTC, AWS: AWS restricted Bedrock on our account. Since 13:41 UTC every call from litellm-proxy fails. All of our Bedrock traffic comes from 152.55.176.0/20 and 208.77.244.0/22. AWS hasn't named a trigger, but the timing points to the same IP reputation issue.
Errors / logs
workflow-worker (production):
2026-10-13 13:05:51 Starting full Salesforce sync ... (background_mode=True)
2026-10-13 13:05:54 Lead sync failed: Salesforce API error: [{"message":"Session expired or invalid","errorCode":"INVALID_SESSION_ID"}]litellm-proxy (production), every Bedrock request:
litellm.BadRequestError: BedrockException - {"message":"Operation not allowed"}Salesforce security email: "This activity was detected originating from a high-risk anonymizing proxy ... IP Address: 152.55.185.5"
Questions
- Are you aware of 152.55.176.0/20 or 208.77.244.0/22 being listed as anonymizing / high-risk by a threat-intel feed? Can you get them reclassified with Salesforce and AWS?
- How do we enable static outbound IPs for workflow-worker, vibesell-api and litellm-proxy? Are they dedicated to our project or shared, and are they included on Pro?
- Will static IPs stay fixed, so we can register them with Salesforce and AWS?
Since April 2026 Salesforce freezes users for any API traffic it rates high-risk, so other Railway users syncing with Salesforce are probably hitting this too.
1 Reply
a month ago
Both ranges are part of our shared egress space. Many independent workloads leave through the same addresses, so a reputation rating on one of them reflects all the traffic behind it, not just yours. We don't pursue reclassification or delisting with third-party providers, Salesforce and AWS included.
On Pro you can enable Static Outbound IPs for each service. Open the service's Settings, turn on Enable Static IPs in the Networking section, check the IPv4 addresses shown for your region, and redeploy. You can also do this from the CLI with railway outbound-network static-ip enable --service <name>. Do this for workflow-worker, vibesell-api and litellm-proxy.
The assigned addresses stay the same unless you move the service to another region, so you can register them with Salesforce and AWS. They aren't guaranteed to be dedicated and may be shared with other customers. Dedicated outbound IPs aren't available on any plan.
Status changed to Awaiting User Response Railway • about 1 month ago
a month ago
This thread has been marked as solved automatically due to a lack of recent activity. Please re-open this thread or create a new one if you require further assistance. Thank you!
Status changed to Solved Railway • about 1 month ago