diff --git a/horde_worker_regen/process_management/process_manager.py b/horde_worker_regen/process_management/process_manager.py
index bfe34f75..d1adb28a 100644
--- a/horde_worker_regen/process_management/process_manager.py
+++ b/horde_worker_regen/process_management/process_manager.py
@@ -702,6 +702,10 @@ def get_process_info_strings(self) -> list[str]:
 
         return info_strings
 
+    def all_waiting_for_job(self) -> bool:
+        """Return true if all processes are waiting for a job."""
+        return all(p.last_process_state == HordeProcessState.WAITING_FOR_JOB for p in self.values())
+
 
 class TorchDeviceInfo(BaseModel):
     """Contains information about a torch device."""
@@ -1577,6 +1581,7 @@ def receive_and_handle_process_messages(self) -> None:
                 break
 
             self._in_deadlock = False
+            self._in_queue_deadlock = False
 
             if isinstance(message, HordeProcessHeartbeatMessage):
                 self._process_map.on_heartbeat(
@@ -1967,7 +1972,7 @@ def preload_models(self) -> bool:
                         f"Already preloading {num_preloading_processes} models, waiting for one to finish before "
                         f"preloading {job.model}",
                     )
-                self._preload_delay_notified = True
+                    self._preload_delay_notified = True
                 return False
 
             self._preload_delay_notified = False
@@ -2061,6 +2066,7 @@ def handle_process_missing(job: ImageGenerateJobPopResponse) -> None:
             if job.model is not None:
                 logger.debug(f"Expiring entry for model {job.model}")
                 self._horde_model_map.expire_entry(job.model)
+
                 try:
                     self.jobs_in_progress.remove(job)
                 except ValueError:
@@ -2112,6 +2118,8 @@ def handle_process_missing(job: ImageGenerateJobPopResponse) -> None:
                     return None
 
         if process_with_model is None:
+            if self._preload_delay_notified:
+                return None
             handle_process_missing(next_job)
             return None
 
@@ -3930,7 +3938,7 @@ def _print_deadlock_info() -> None:
             not self._in_queue_deadlock
             and (self._process_map.num_busy_processes() == 0 and len(self.job_deque) > 0)
             and len(self.jobs_in_progress) == 0
-        ):
+        ) or (self._process_map.all_waiting_for_job() and len(self.job_deque) > 0):
 
             currently_loaded_models = set()
             model_process_map: dict[str, int] = {}
@@ -3945,13 +3953,34 @@ def _print_deadlock_info() -> None:
                     self._last_queue_deadlock_detected_time = time.time()
                     self._queue_deadlock_model = job.model
                     self._queue_deadlock_process_id = model_process_map[job.model]
+                    break
+            else:
+                logger.debug("Queue deadlock detected without a model causing it.")
+                _print_deadlock_info()
+                self._in_queue_deadlock = True
+                self._last_queue_deadlock_detected_time = time.time()
+                # we're going to fall back to the next model in the deque
+                self._queue_deadlock_model = self.job_deque[0].model
 
         elif self._in_queue_deadlock and (self._last_queue_deadlock_detected_time + 10) < time.time():
             logger.debug("Queue deadlock detected")
             _print_deadlock_info()
-            logger.debug(f"Model causing deadlock: {self._queue_deadlock_model}")
-            if self._queue_deadlock_process_id is not None:
-                self._replace_inference_process(self._process_map[self._queue_deadlock_process_id])
+            if self._queue_deadlock_model is not None:
+                logger.debug(f"Model causing deadlock: {self._queue_deadlock_model}")
+                if self._queue_deadlock_process_id is not None:
+                    self._horde_model_map.expire_entry(self._queue_deadlock_model)
+                    self._replace_inference_process(self._process_map[self._queue_deadlock_process_id])
+            else:
+                logger.warning("Queue deadlock detected but no model causing it.")
+                num_processes_replaced = 0
+                # We're going to replace up to two process which aren't busy
+                for process in self._process_map.values():
+                    if not process.is_process_busy():
+                        self._replace_inference_process(process)
+                        num_processes_replaced += 1
+                    if num_processes_replaced >= 2:
+                        break
+
             self._in_queue_deadlock = False
             self._queue_deadlock_model = None
             self._queue_deadlock_process_id = None