Hey there,
I’ve seen something similar happen when the job counter inside the temporary dispatch buffer doesn’t handshake properly with the flow scheduler index. When that happens, the job looks “queued” forever even though the pipeline technically considers it “pending on handshake.” It’s weird but harmless until the internal token gets recycled.
Try this:
Go into your Flow Preferences > Advanced, and toggle the “Evaluate on Submission” option twice (off → on → off). That forces a refresh of the queue references.
Then clear the Dispatch Cache folder (not the main job folder — just the one under the service temp). That often reinitializes the stalled listener thread.
Restart the Watchdog, but don’t restart the Flows service right away — give it about 45 seconds so the shadow instance can unregister.
If it still shows “queued,” check whether your action list checksum matches the flow definition checksum (they’re listed in the metadata XML). A mismatch can sometimes trick Flows into waiting for a phantom dependency.
It sounds a bit abstract, but usually once the dispatcher rebuilds its state file, jobs start moving instantly.
Hope this nudges it loose!