a month ago
Project: dirtyline-manager
Service: dirtyline-manager-production
Builder: builder-rjjayj (Metal builder)
Every build attempt is failing at the same step: "load metadata for docker.io/library/node:20-alpine" with a network timeout error: failed to solve: DeadlineExceeded: node:20-alpine: failed to resolve source metadata for docker.io/library/node:20-alpine: failed to do request: Head "https://registry-1.docker.io/v2/library/node/manifests/20-alpine": dial tcp :443: i/o timeout
This has happened across multiple attempts over the last ~20 minutes, with different resolved IPs each time (100.58.88.94, 23.20.232.38, 184.192.139.247) and different image tags (node:20-alpine, node:20-alpine3.20). I also tried triggering a fresh build via an empty commit to get a different builder assignment, but the same builder (builder-rjjayj) and same error persisted.
Railway status page shows all green, so this looks like an issue specific to this builder's network path to Docker Hub. Could you check builder "builder-rjjayj" or reassign my build to a different builder? Failed deployment IDs: 9ad364b1, 084458b7
4 Replies
Status changed to Awaiting Railway Response Railway • about 1 month ago
a month ago
We are seeing the same failure on a separate Railway project and separate Metal builder.
Our builder is builder-osnvon; your report references builder-rjjayj.
We have had four consecutive deployments of the exact same Git commit fail during Build › Build image, before npm install or our application build begins.
The error is consistently:
failed to solve: DeadlineExceeded: failed to load cache key: failed to do request: Head "https://registry-1.docker.io/v2/library/alpine/manifests/latest": dial tcp :443: i/o timeout
Failed deployment IDs:
c5174460
ba963797
75b7629c
be74e658-4fcf-40e1-9aeb-e0da5f2bd98d
One of our failed attempts also resolved Docker Hub to 100.58.88.94, the same IP reported in this thread.
Our project is using Railpack on the Metal builder. The Railpack plan is generated successfully, but BuildKit fails before application build execution. Our repository contains no Alpine reference in the Railway build path, and the generated Railpack images shown in the logs are Railway GHCR images.
Existing production remains healthy on the previous successful deployment, so this appears isolated to the builder egress/build-image path rather than application runtime.
Could Railway please investigate outbound connectivity from Metal builders to registry-1.docker.io:443?
Builder: builder-osnvon
Region: EU West / Amsterdam
Railpack: Default
Latest failed deployment: be74e658-4fcf-40e1-9aeb-e0da5f2bd98d
a month ago
We're seeing the exact same issue on a separate project , also on builder-rjjayj.
Error 1: docker.io/docker/dockerfile:1 → dial tcp 32.196.9.70:443: i/o timeout
Error 2: docker.io/library/ruby:3.3.11-slim → dial tcp 98.85.115.1:443: i/o timeout
Same builder, different images, different IPs — consistent with the pattern reported here. Production build of the same repo succeeds, so this is isolated to this builder's egress path, not our Dockerfile.
a month ago
same issue for my project too
a month ago
This was a problem on our end, not your Dockerfiles. A couple of our Amsterdam build machines lost their network path to Docker Hub, so anything that landed on them failed at the base image step. It's resolved and builds have been clean since 05:47 UTC. Sorry about that. If it comes back, let us know.
Status changed to Awaiting User Response Railway • about 1 month ago
Status changed to Solved yudiarnanda • about 1 month ago