Closed lmoesle closed 1 year ago
@lmoesle Would not do that - to complicated and depends on the engine we use. In my Point of View we should just publish the other artifacts again.
@dominikhorn93 What other artifacts do you mean? Should we just ignore the failed deployments and inform the user that their artifact is not valid?
It can be complicated to do a "revert" automatically. Maybe we did override a json schema and don't know which one was deployed before? I think we should just inform the user and show which artifacts fail. In my point of view that is an ongoing "try and error" task until everything succeeds.
If a prod deployment fails the user should redeploy the last successful release.
Sounds good to me. The deployment proxy already implements this behavior. Therefore I will close this issue.
Description
If I deploy multiple files to the digiwf-engine (via deployment-proxy) and some deployments fail and some succeed, I would like to revert the successful deployments to avoid a broken state.
Sub-Tasks
Note: We should wait until a revert deployment api is created in digiwf-core.