24 days ago
Our healthy application needs a controlled variable update while retaining a selected exact source commit. No deployment or variable change has been made to test this question.
The native Railway connector exposed to our assistant has an operation described as:
redeploy(projectId, environmentId, serviceId, deploymentId?)
Its description says it reuses an existing build. The rollback guide describes Redeploy using original source and original variables, while Railway employee sam-a's Dashboard answer describes rebuilding the selected source/config with current service variables.
For THIS native connector operation, please clarify:
- Does it reuse the selected build or rebuild its source? Which Railway platform operation does it call?
- Does its new runtime receive CURRENT COMMITTED service variables or the selected deployment's HISTORICAL variable snapshot? Are still-STAGED variable changes consumed or left pending?
- What supported sequence retains the selected exact source commit while activating one newly committed variable, with a nonsecret check of source/config binding before activation? If this connector cannot guarantee that sequence, please state that limitation.
Conflicting descriptions:
- https://docs.railway.com/guides/roll-back-bad-deploy
- https://docs.railway.com/deployments/deployment-actions
- https://docs.railway.com/cli/redeploy
- https://station.railway.com/questions/reuse-a-previous-deployment-image-while-6d44a70d
This is a product/adapter documentation question. No current outage is claimed. We are requesting an explanation only: please do not deploy, restart, alter variables, run SQL or change any resource. Any suggested operation will be reviewed separately.
Prepared with Railway's support template (v1).
1 Reply
Status changed to Awaiting Railway Response Railway • 24 days ago
24 days ago
We can't vouch for how your assistant's connector maps its redeploy operation onto our platform, so we can't confirm its "reuses an existing build" description. What follows describes our own Redeploy and Rollback actions.
Redeploy creates a new deployment from the selected deployment's original source and build/deploy configuration, so it can rebuild old code even after that deployment's image is gone. Rollback is the action that restores the selected deployment's Docker image together with its custom variables, and it only appears while the image is inside your plan's retention window.
On variables, Redeploy is the documented way to pick up variable changes (railway redeploy), so it runs with the current committed variables, not the old snapshot. Rollback is the one that restores the old variable set. Staged edits aren't live until they're committed. "Deploy" in the staged-changes banner commits them and redeploys the affected services. Holding Alt while clicking Deploy, or calling environmentPatchCommitStaged, commits them without a redeploy (Staged Changes).
So, to keep an exact commit and add one new variable: stage the variable, commit it without redeploying, review the diff and the selected deployment's source and configuration in the dashboard, then Redeploy that specific deployment. That deployment is rebuilt from its original source rather than reused from its old image, so you can't expect a byte-identical image. If the connector really reuses images, it can't guarantee that it picks up the new variable.
Status changed to Awaiting User Response brody • 24 days ago
17 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 • 17 days ago