Determine if a health check status option of inconclusive is really better than fail

A full deployment has taken about 3 hours to complete if all goes well. During the time before the deployment has completed successfully, many health checks will run and fail. Is it worth having these tests have an inconclusive status option (which they currently do) or is it acceptable to have that test fail instead. Keep in mind that these health checks will re run on a timer that could be set to any interval. Do we (or will we, or could we) treat inconclusive results differently than failures (in automation terms)?