We’re seeing Railway environment staged-change workflows get stuck indefinitely in APPLYING for production.

Project ID: 0c92aacc-553a-4b77-b944-f276f9b7eafd

Environment ID: 605e5949-b246-4396-8995-5d746de0e913

Symptoms:

• Dashboard shows “Applying N changes” for services, including services that have since been deleted.
• Deleted services still appear to have pending apply workflows tied to them.
• While these workflows are stuck, we cannot reliably apply variables or shared variables from the dashboard/API.
• Public GraphQL variable/staged patch APIs appear to enqueue commitChanges workflows that remain stuck.
• We found internal dashboard links referencing Temporal workflows like: commitChanges/605e5949-b246-4396-8995-5d746de0e913...
• Manual service creation/deploy via dashboard-style service create + serviceInstanceUpdate + serviceInstanceDeployV2 can work, but staged environment patch commits appear wedged.

Can you clear/cancel the stuck commitChanges workflows for this environment, and confirm whether environmentPatchCommitStaged / variableCollectionUpsert can leave an environment permanently stuck in APPLYING when target services are deleted? We also need guidance on the correct API sequence for creating a Docker image service with start command + variables without leaving pending staged changes.