Skip to content
Contract clarification: pause one Production Cron without deployment or consuming unrelated staged changes
trisfas
OP

a month ago

Hello Railway Support,

We need guidance only. Please do not change any resources or configuration on our behalf.

In project diplomatic-inspiration, the Production service proposalproof-backup-cron has a native Railway cron schedule of 0 2 * * * (UTC). Its latest deployment is SUCCESS with its container stopped, but the UI and service metadata still show the next scheduled run. Our application-level backup and TLS-only approval flags are both disabled; the previous temporary TLS authorization has expired.

Production also has six unrelated staged changes that would add soothing-wholeness and its GitHub/build/start configuration. These changes must remain untouched and unapplied.

Please explicitly answer:

  1. Does serviceInstanceUpdate(serviceId, environmentId, input: {cronSchedule: null}) immediately stop future native Cron invocations of the existing deployment, or only change configuration for a future deployment? How are already queued invocations handled?
  2. Does this API operation create a staged change, require Apply/Commit, automatically create a deployment, restart the container, or execute any start/pre-deploy command? Is removing the existing deployment required?
  3. With unrelated staged changes already present, does this operation leave their exact contents and unapplied status unchanged? If a commit is required, is there a supported way to commit only this service's schedule change without redeploying or consuming the other changes? Does Alt+Deploy commit the entire changeset?
  4. Can deployment-level railway.json/railway.toml settings override a null service schedule now or re-enable it on a later deployment? Which layer is authoritative for the existing deployment's scheduler?
  5. Is environmentId guaranteed to restrict this operation to Production? The live schema's experimental fork-related description appears broader than the per-environment wording in the API guide.
  6. What exact supported UI/API operation satisfies all these constraints, and which read-only fields conclusively verify that scheduling has stopped? Please distinguish committed service settings, deployment snapshots and the scheduler's actual state, including expected propagation delay.

We cannot authorize a build, deployment, restart, workload execution, source change, resource deletion, or application/discarding of the unrelated staged changes. If no method provides these guarantees, please state that explicitly and explain the minimal additional operation that would be required.

Thank you.

$10 Bounty

1 Reply

Status changed to Awaiting Railway Response Railway • about 1 month ago


Railway
BOT

a month ago

This thread has been opened as a bounty so the community can help solve it.

Status changed to Open Railway • about 1 month ago


lancetmason

a month ago

Hey @trisfas, I think I can help with this. Do you have a github repo? Would you be open to adding me as a contributor or doing a screenshare?


Welcome!

Sign in to your Railway account to join the conversation.

Loading...