The three builds currently running (dashboard, storefront, api-worker-trial) started at 15:30 UTC in us-east4, and on dashboard the turbo build ran with 0 of 3 tasks cached (about 7 minutes of compile), followed by several minutes each copying node_modules and exporting the image, so most of the extra time is the full uncached rebuild plus a large image. Per our slow deployments guide, cache misses and large images are the main drivers of long builds, so trimming dependencies (especially dev dependencies in node_modules) and keeping the final image small will shorten both the build and the image upload and pull.

same. why does railway have these issues so often?