a month ago
We have three guarded deployment attempts failing before any build starts. The previous successful deployment remains running unchanged.
Railway CLI version: 5.52.1. Command: railway up --detach. The CLI exits with code 1 without returning a deployment ID; subsequent deployment read-back shows FAILED at SNAPSHOT_CODE, with no build or image.
The latest attempt timed out after approximately 91.3 seconds. The error shown is: “Failed to create code snapshot. Please review your last commit, or try again. If this error persists, please reach out to the Railway team.”
We have not established the cause or measured the transmitted archive size. An incomplete/interrupted upload is only a hypothesis, not a confirmed Railway outage.
What diagnostic or recovery steps do you recommend to establish whether the source upload completed and why snapshot creation failed? We can supply deployment identifiers privately through an appropriate Railway support channel.
No customer data, source code, raw logs, credentials or private identifiers are included in this public request.
1 Reply
a month ago
A source upload from railway up has to finish inside Cloudflare's 125-second request timeout. If the complete compressed archive can't reach us within that window, the deployment fails at SNAPSHOT_CODE with "Failed to create code snapshot". No build starts, so you won't get build logs or an image.
To check whether the upload is the problem, measure the actual upload duration and throughput from the machine or CI runner that runs railway up. A deploy that worked before, or an archive that packages fine locally, doesn't show that the current archive can upload within the window. Archive size can still be the cause without a 413 error, because a payload under any byte-size limit can still take too long to upload on a given connection.
To recover, add or update a .railwayignore that excludes anything not needed to build or run the app, such as local caches, logs, datasets, virtual environments and generated output. Keep your source, manifests, lockfiles and build inputs. Raising RAILWAY_HTTP_TIMEOUT only changes how long the CLI waits. It doesn't extend Cloudflare's timeout. If nothing more can safely be excluded, you'll need to deploy from a connection that can upload the archive within the window.
Status changed to Awaiting User Response Railway • about 1 month ago
24 days ago
This thread has been marked as solved automatically due to a lack of recent activity. Please re-open this thread or create a new one if you require further assistance. Thank you!
Status changed to Solved Railway • 24 days ago