My deploy for service earnest-miracle (project: Aventuras Jorge, repo: aventurasjorge/aventurasjorge.com, URL: https://railway.com/project/ed82e5d6-6af7-45c4-b6fb-d0e80bc0da8a) has failed twice in a row, both times scheduled on the same builder and failing at the identical step:

[inf] scheduling build on Metal builder "builder-yxigcz"

[inf] fetched snapshot ... (201 kB bytes)

[inf] fetching snapshot

[inf] unpacking archive

[err] Build Failed: fsutil.NewFS(/var/lib/builder/build-sessions/.../snapshot-target-unpack/upload_to_github): resolve : lstat .../snapshot-target-unpack/upload_to_github: no such file or directory

First occurrence: 2026-09-06T18:18:47Z, build session 1e890952-8553-4fc6-b5d7-2a6f94dd8d37

Second occurrence: 2026-09-06T18:24:52Z, build session ac2243ec-7094-439c-9976-c37eb553920f

Both attempts were scheduled on the same builder (builder-yxigcz) and failed at the exact same internal path (snapshot-target-unpack/upload_to_github) immediately after the snapshot was unpacked — before my Dockerfile (Dockerfile.api) ever started running. The snapshot itself is small (201 kB), so this doesn't look like a repo-size issue.

Since it's reproduced twice on the same builder rather than appearing as a one-off, I suspect there may be corrupted/stuck state on that specific builder node rather than something on my end. Could you take a look at builder-yxigcz specifically, or let me know if there's a way to force my next build onto a different builder while this is investigated?

Happy to provide the full build logs or repo details if useful.