a month ago
Does Railway support infrastructure-enforced, default-deny outbound network policies for an isolated test project, including IPv4, IPv6, and DNS restrictions?
Specifically, can a service be restricted to communicating only with explicitly authorized services within its own project, while preventing access to all external public endpoints?
If supported, what configuration and independent verification methods are available?
We are evaluating the architecture only and are not requesting provisioning or deployment.
1 Reply
Status changed to Awaiting Railway Response Railway • about 1 month ago
a month ago
We don't offer a customer-configurable default-deny egress policy for services. Every service gets outbound IPv4 internet access, and there are no settings for destination allowlists, egress firewall rules, or restricting which external DNS names a service can resolve.
What you can control: outbound IPv6 is off by default and enabled per service, and while it's off, IPv6 connections fail with ENETUNREACH (Outbound Networking). Private networks are isolated per project and environment at the infrastructure level, so services in other projects or environments can't reach yours over the private network. Within a single environment, though, all services share one trust boundary, and there's no per-service allowlist inside it (How Private Networking Works).
If you want to verify this yourself, you can SSH into a service with the CLI and test outbound connections directly. Penetration testing your own services needs no pre-approval, but our standard rate limits and DDoS protections still apply.
Status changed to Awaiting User Response sam-a • 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