Failures that are Kuppit's
Some failures are not in your repository and not on your Variables tab. The page says This is Kuppit's to fix, not something in your repository, and it means it: nothing you change will make the next attempt succeed unless the cause was transient. What you can do is Redeploy once, and if it fails the same way, wait. Whatever was serving keeps serving throughout.
These entries say what each message is, so that you know whether waiting is the right call.
Kuppit encountered an internal build error
The build infrastructure rejected the build request itself, before your code ran: a builder that was not pinned, a request the infrastructure would not accept. Nothing in your repository contributes to it. Redeploy once; if it repeats, it is Kuppit's to repair.
Kuppit lost track of the build
Kuppit started a build and could no longer find it when it went to check. It is a transient fault on Kuppit's side and a redeploy usually succeeds.
Traffic did not move to the new revision
The revision started and passed its health check, but production traffic did not reach it in time. The previous revision kept serving. Redeploy; the second attempt normally converges.
The deployment stopped making progress
Kuppit retried a step repeatedly without it completing and gave up rather than leave the deployment in flight forever, which would block every later deployment of the service. Redeploy. If it happens again, it is being looked at.
A resource Kuppit expected was missing
Something Kuppit had recorded, a build, an image, a revision, was not there when Kuppit went back for it. Redeploy once. A repeat means Kuppit's record and the infrastructure have disagreed, which is Kuppit's to reconcile.
Kuppit was rate limited while deploying
The infrastructure throttled Kuppit's requests. Wait a few minutes and redeploy.
The referenced database failed to provision
A variable references a database whose provisioning ended in failure. The database's own page says so and offers a retry; a deployment cannot proceed against it. Retry the database from its page, or delete it and create another, then redeploy the service.
Kuppit does not yet know where the referenced database is reachable
The database is provisioned but Kuppit has not yet observed the address to connect to. It normally clears within minutes; redeploy then. If the database's card is Ready and this persists, it is Kuppit's.
A variable references a database in another environment
A reference can only target a resource in the same environment, and Kuppit's dashboard does not let you create one that crosses. Seeing this means Kuppit let something through that it should not have. Remove the variable and reference a database in this environment; the crossing reference is a bug worth reporting.
This service cannot reach the database it references
The service and the database are placed on infrastructure that cannot be connected to each other. It is a placement fault, not a configuration one, and nothing on the Variables tab changes it.
The deployment did not reach production
Kuppit has no more specific words: the deployment ended without a recorded cause. Redeploy, and read Advanced details for anything the infrastructure said.
Kuppit encountered an unexpected error
A failure Kuppit has no sentence for. Kuppit treats it as its own on purpose, because guessing that your application caused it is the more expensive mistake. Redeploy once; Advanced details holds the raw code and message.