Answer
A message that fails to reach its endpoint moves to Waiting so it can automatically retry, and how long it stays there before either succeeding or ultimately failing is governed by the max-retries and retry-interval settings on the relevant REST sender/routing address — manually restarting it doesn't change that underlying retry schedule. If the endpoint genuinely can't be reached, no amount of manual retrying will help; confirm first that the target endpoint is valid and actually reachable from the middle tier's own network path (not just from your own machine) before assuming a message-queue or connector configuration problem.