Three consecutive builds for the health-edge backend service have stalled on the same step:

load metadata for docker.io/library/node:22.12-slim

Project: HealthEdge

Project ID: eb7b20cb-5fe3-4de3-951f-ec6fd6e86b20

Service: health-edge (a738f1ee-17b7-4ac2-82f4-a906ab9def8f)

Environment: production

Region: us-west2

Metal builder: production-builderv3-us-west1-0hwp

Timeline (last 90 minutes):

Build 1 (deployment cf34ed10) — 32 minutes on this step before I aborted

Build 2 (after railway redeploy via CLI) — failed at "Build > Build image" after 20 minutes

Build 3 (deployment 261a0d09, currently active) — 12+ minutes on the same step at time of writing

The service builds from a Dockerfile in a monorepo. Dockerfile starts with FROM node:22.12-slim AS builder. Same Dockerfile shape was building cleanly through the end of last week (last successful deployment 2026-04-30T10:14 UTC).

Earlier today I changed the service Root Directory in the dashboard from /backend to / so the build context covers the whole monorepo (needed for an npm workspace package extracted earlier). railway.toml is at repo root with:

[build]

builder = "DOCKERFILE"

dockerfilePath = "backend/Dockerfile"

That config change took effect — Railway is reading the new Dockerfile and starting the build correctly. The stall is consistently at the docker.io/library/node:22.12-slim metadata fetch on the Metal builder.

Is builder production-builderv3-us-west1-0hwp or the metro region's Docker Hub egress having issues? Production has been frozen on a stale image for ~22 hours because of this.

Public-facing health endpoint stays healthy throughout (old image is still serving): https://health-edge-production.up.railway.app/api/health returns 200, which is masking the deploy backlog.

Happy to provide deployment IDs, build logs, or repo access.