a month ago
Hello Railway team,
Before a future controlled migration, I need to verify exact recovery of a known-good Dockerfile-built deployment. This is a platform rollback question, not a request to debug application code. No deployment or rollback has been attempted for this validation.
Can you confirm whether dashboard Rollback, after a failed replacement deployment, restores the exact previous image SHA-256 digest without rebuilding, together with its deployment configuration and environment variables? Please distinguish this from Redeploy.
Please clarify:
- The image retention window applicable to our current account/plan, and when that window starts. If account-specific details cannot be discussed publicly, please provide a supported secure route to confirm them.
- Which settings, shared/referenced variables, secrets, networking settings or persistent resources are NOT restored by Rollback. Please do not disclose any secret values.
- How the exact known-good image and required configuration can be preserved for recovery beyond the retention window, without depending on a rebuild or floating base-image tag.
- What read-only evidence we can obtain now to prove that exact rollback will remain available after a future failed replacement.
Project/deployment identifiers and image digest are intentionally omitted from this public thread. They can be provided through a supported private channel if necessary.
Please do not modify, redeploy, restart, restore or otherwise change any live service. We are requesting confirmation and instructions only.
Thank you.
1 Reply
a month ago
Rollback restores the selected deployment's previous image, settings and variables as a new deployment, without rebuilding. Redeploy is different: once an image is past retention, Redeploy rebuilds from the original source code using that deployment's original variables. So a Dockerfile with floating base tags could produce a different image.
Image retention depends on the plan: 24 hours on Free/Trial, 72 hours on Hobby, 120 hours on Pro and 360 hours on Enterprise. The window starts when the deployment is removed, and the image is kept for that long after removal. Once a deployment is outside the window, the Rollback option no longer appears in its menu (image retention policy, deployment actions).
Each deployment's Details tab keeps a snapshot of the service's own variables, and it stays readable after Rollback is no longer offered. That snapshot leaves out shared/project variables and Railway-provided variables. Reference variables are stored as the reference text, not the resolved value, and sealed variables appear blank. Rollback restores the image and that deployment's variables. Persistent volumes are separate resources and are not rolled back.
There is no read-only way to guarantee Rollback ahead of time. The evidence you can see is whether the Rollback option shows on a previous deployment, plus that deployment's Details tab. Beyond the window, the reliable way to get recovery that doesn't depend on a rebuild is to deploy the service from a pinned image digest in a registry you control.
Status changed to Awaiting User Response Railway • about 1 month ago
a month 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 • about 1 month ago