a year ago
I'm very excited to see multi-region, as it can save me time and effort in deploying load balancing.
However, I want to understand some details and hope the team can clarify for me.
My project is a small microservice that uses Caddy as a reverse proxy to multiple sub-applications (there's an image attached).
For example: User ----> Caddy --foo.railway.internal--> foo-service
I've set up multi-region for my Caddy, deploying nodes in Singapore and the US West. According to the article at https://railway.app/changelog/2024-12-06-multi-region-replicas, I'm sure that User ----> Caddy can access the nearest node, but
I'm not sure if Caddy --foo.railway.internal--> foo-service access is also nearest.
For example: Can a user in Singapore access both the Singapore Caddy and the Singapore foo-service, or is that still uncertain?
3 Replies
a year ago
Hello,
Routing traffic to the nearest node is only available for public traffic.
If you make a request to a service with multi-region replicas over the private network, you could end up making the request to a deployment in any region.
Status changed to Awaiting User Response Railway • over 1 year ago
a year ago
Thank you for your response. For backend services, this is a common requirement. Are there any plans to handle this at the private network level in the future?
Or perhaps multi-region could provide multiple Private Networks, so I can manually configure them in Caddy?
Status changed to Awaiting Railway Response Railway • over 1 year ago
a year ago
That would be a nice enhancement, can you raise that in https://help.railway.com/feedback?
In the meantime, going to mark this thread as solved.
Status changed to Awaiting User Response Railway • over 1 year ago
Status changed to Solved brody • over 1 year ago