Try to redeploy your service by clicking on it, then press CTRL + K, then select deploy latest commit. You can also try settings NO_CACHE=1 in your service variables, to rule out any caching issues.

Your error means one of these:

The Dockerfile uses BuildKit cache mounts that Render/Railway doesn't support in that form.

The Dockerfile was written for another builder (Depot, Docker Buildx, GitHub Actions, etc.).

The deployment platform has stricter Dockerfile validation.

The correct fix

You need to edit the Dockerfile, specifically around line 235, and either:

Remove the --mount=type=cache option, or

Change it to the syntax your platform expects.

I can fix it for you.

Please send one of these:

The GitHub repository URL for the project (preferred), or

The Dockerfile (or at least lines 220–245).

Thanks for looking into this, but I don't think this matches my issue — this project doesn't have a Dockerfile at all. It's a static site that Railway builds automatically via Railpack's Staticfile detection, using a Caddyfile for routing/headers. There's no Docker build step, no BuildKit cache mounts, nothing at "line 235" for me to send you.

Also, the failures aren't happening at the build stage — they're failing during the upload step of railway up, before any build even starts:

Indexing...

Uploading...

error sending request for url (https://backboard.railway.com/project/41a3028c-012a-48b1-9d6a-6801baed4619/environment/886cfd6c-2f0d-4ae2-b7a5-6c01dda698d3/up?serviceId=58d4118b-2010-41b7-8c27-c76c59c366cc)

Caused by:

operation timed out

This has happened consistently across ~8 attempts over the past 24 hours. I also tried setting NO_CACHE=1 as suggested earlier — same timeout, since it's failing before the build/cache would even come into play.

Could you check the upload/API infrastructure for:

Project ID: 41a3028c-012a-48b1-9d6a-6801baed4619

Service ID: 58d4118b-2010-41b7-8c27-c76c59c366cc

Region: US East

One deploy did succeed yesterday (2026-07-19 22:41 UTC+3) — deployment ID 8eaa206a-1dde-4713-86e9-ecb3b34ff1e0 — so the service itself is configured correctly. Everything since then has failed the same way.

Update: I also tried the dashboard method — pressing CTRL+K and selecting "Deploy latest commit" — and got this error:

Cannot redeploy without a snapshot

This confirms the root cause: the CLI's railway up command can't create a snapshot because the upload to backboard.railway.com is timing out. Without a snapshot, neither the CLI nor the dashboard can deploy anything.

So the issue isn't about Dockerfile or BuildKit — it's that the snapshot upload API endpoint is rejecting/timing out all requests.

This has happened consistently across ~8 attempts over 24 hours:

Last successful deploy: 2026-07-19 22:41 UTC+3 (deployment ID: 8eaa206a-1dde-4713-86e9-ecb3b34ff1e0)

All attempts since then fail at the upload step with operation timed out

Project details:

Project ID: 41a3028c-012a-48b1-9d6a-6801baed4619 (energetic-light)

Service ID: 58d4118b-2010-41b7-8c27-c76c59c366cc

Region: US East

Environment: production

Could you check the backboard.railway.com upload API infrastructure? It appears to be down or experiencing issues for this account/project.

I've also set NO_CACHE=1 as suggested in the earlier response, but that didn't resolve the upload timeout since it fails before the build even starts.

I've looked at some old threads with similar issues, it's probably something wrong on railway side, and is solved by railway team, so I think nothing we can do on our end, or you can try deploy it as a new service

Threads I found:

https://station.railway.com/questions/build-failing-on-snapshot-d2f879fd

https://station.railway.com/questions/failed-to-create-code-snapshot-please-r-9f020901

see medim's answer below

Your project is too big and or you're having network issues, railway up has a timeout window of 15 minutes.

Make sure large files, build artifacts, databases, backups, and dependency directories are excluded via .gitignore / .railwayignore.

You can also run the railway up with the --verbose tag to see if it produces any extra logs.

nothing is working

Thanks for the suggestion! I investigated and here's what I found:

Size isn't the issue: I created a .railwayignore file to exclude brochure assets and cache directories, but the upload still fails at the exact same point. The 48MB that's being uploaded is all legitimate website assets (images, PNGs, JPGs, WebPs) that are committed to git and needed for the live site — that should upload fine within a 15-minute window.

The real problem: The timeout is happening at the upload step to backboard.railway.com, not during indexing or the build phase. This is consistent across ~9 attempts over 24 hours, always failing on the same API endpoint.

Context:

Last successful deploy: 2026-07-19 22:41 UTC+3 (deployment ID: 8eaa206a-1dde-4713-86e9-ecb3b34ff1e0)

All attempts since then fail with the same operation timed out error on backboard.railway.com

Project size (48MB) shouldn't be an issue given the 15-minute timeout window

I also set NO_CACHE=1 as suggested earlier — same timeout

Thanks for checking. This is a static HTML/CSS/JavaScript site with no build system or dependencies. The only "framework" is a Caddyfile for routing and cache headers on Railway's Railpack.

No Node, Python, Go, or other runtime — just vanilla static files. The repo is 36.2MB (all HTML, CSS, JS, and image assets), and that's after removing duplicate files.

The upload consistently times out at the backboard.railway.com API even though it's under the 45MB limit, so it doesn't appear to be a language/framework issue — more likely an API-side connectivity problem.

Try optimizing your image assets, have you tried running railway up --verbose as I stated earlier?

I'm sure this is not an API connectivity problem, but can you also try doing railway up through other network?