diff --git a/arangod/Pregel/Worker.cpp b/arangod/Pregel/Worker.cpp
index 8586e4fb2d..f137176955 100644
--- a/arangod/Pregel/Worker.cpp
+++ b/arangod/Pregel/Worker.cpp
@@ -255,7 +255,6 @@ void Worker<V, E, M>::receivedMessages(VPackSlice const& data) {
 
     // Trigger the processing of vertices
     if (_config.asynchronousMode() && _state == WorkerState::IDLE) {
-      MUTEX_LOCKER(guard, _commandMutex);
       _continueAsync();
     }
   } else if (_config.asynchronousMode() &&
@@ -548,7 +547,6 @@ void Worker<V, E, M>::_finishedProcessing() {
       }
     }
     if (proceed) {
-      MUTEX_LOCKER(guard, _commandMutex);
       _continueAsync();
     }
   } else {  // no answer expected
@@ -561,12 +559,16 @@ void Worker<V, E, M>::_finishedProcessing() {
 /// in async mode checks if there are messages to process
 template <typename V, typename E, typename M>
 void Worker<V, E, M>::_continueAsync() {
-  uint64_t cnt = _writeCache->containedMessageCount();
-  if (_state == WorkerState::IDLE && cnt > 0) {
+  MUTEX_LOCKER(guard, _commandMutex);
+  if (_state != WorkerState::IDLE || _writeCache->containedMessageCount() == 0) {
+    return;
+  }
     // avoid calling this method accidentially
-    _state = WorkerState::COMPUTING;
-    
-    if (cnt < _messageBatchSize) {
+  _state = WorkerState::COMPUTING;
+  
+  ThreadPool *pool = PregelFeature::instance()->threadPool();
+  pool->enqueue([this]{
+    if (_writeCache->containedMessageCount() < _messageBatchSize) {
       usleep(50000);
     }
     {  // swap these pointers atomically
@@ -576,12 +578,13 @@ void Worker<V, E, M>::_continueAsync() {
         _requestedNextGSS = true;
       }
     }
+    MUTEX_LOCKER(guard, _commandMutex);
     // overwrite conductor values with local values
     _conductorAggregators->resetValues();
     _conductorAggregators->aggregateValues(*_workerAggregators.get());
     _workerAggregators->resetValues();
     _startProcessing();
-  }
+  });
 }
 
 template <typename V, typename E, typename M>