diff --git a/inspirehep/config.py b/inspirehep/config.py
index cef268e301..a33d9026ef 100644
--- a/inspirehep/config.py
+++ b/inspirehep/config.py
@@ -58,7 +58,9 @@
 """This feature flag will prevent to send a ``replace`` update to legacy."""
 FEATURE_FLAG_USE_ROOT_TABLE_ON_HEP = False
 FEATURE_FLAG_ENABLE_SNOW = False
-FEATURE_FLAG_ENABLE_SAVE_WORKFLOW_ON_DOWNLOAD_DOCUMENTS = True
+# This flag is to help with empty key documents from the merge process
+FEATURE_FLAG_ENABLE_DELETE_EMPTY_KEY_DOCUMENTS = True
+
 # Default language and timezone
 # =============================
 BABEL_DEFAULT_LANGUAGE = 'en'
diff --git a/inspirehep/modules/workflows/config.py b/inspirehep/modules/workflows/config.py
index 1bbae1306e..c900f8434e 100644
--- a/inspirehep/modules/workflows/config.py
+++ b/inspirehep/modules/workflows/config.py
@@ -31,3 +31,5 @@
 WORKFLOWS_PLOTEXTRACT_TIMEOUT = 5 * 60
 """Time in seconds a plotextract task is allowed to run before it is killed."""
 WORKFLOWS_MAX_AUTHORS_COUNT_FOR_GROBID_EXTRACTION = 50
+WORKFLOWS_DOWNLOAD_DOCUMENT_TIMEOUT = 5 * 60
+"""Time in seconds a download document task is allowed to run before it is killed."""
diff --git a/inspirehep/modules/workflows/tasks/actions.py b/inspirehep/modules/workflows/tasks/actions.py
index a1d2ff259e..e006d67713 100644
--- a/inspirehep/modules/workflows/tasks/actions.py
+++ b/inspirehep/modules/workflows/tasks/actions.py
@@ -89,7 +89,9 @@
     get_validation_errors,
     log_workflows_action,
     with_debug_logging, check_mark, set_mark, get_mark, get_record_from_hep,
-    delete_empty_key
+    delete_empty_key,
+    timeout_with_config,
+    ignore_timeout_error,
 )
 from inspirehep.modules.workflows.utils.grobid_authors_parser import GrobidAuthors
 from inspirehep.utils.url import is_pdf_link
@@ -401,6 +403,8 @@ def fix_submission_number(obj, eng):
 
 
 @with_debug_logging
+@ignore_timeout_error(return_value=[])
+@timeout_with_config('WORKFLOWS_DOWNLOAD_DOCUMENT_TIMEOUT')
 def populate_submission_document(obj, eng):
     submission_pdf = obj.extra_data.get('submission_pdf')
     if submission_pdf and is_pdf_link(submission_pdf):
@@ -422,15 +426,15 @@ def populate_submission_document(obj, eng):
         LOGGER.info('Workflow data updated with %s new documents' % len(obj.data.get('documents', [])))
     else:
         LOGGER.info('Submission document not found or in an incorrect format (%s)', submission_pdf)
-    delete_empty_key(obj, 'documents')
-
-    if current_app.config['FEATURE_FLAG_ENABLE_SAVE_WORKFLOW_ON_DOWNLOAD_DOCUMENTS']:
-        save_workflow(obj, eng)
-    else:
-        obj.save()
+    if current_app.config['FEATURE_FLAG_ENABLE_DELETE_EMPTY_KEY_DOCUMENTS']:
+        delete_empty_key(obj, 'documents')
+    save_workflow(obj, eng)
+    return
 
 
 @with_debug_logging
+@ignore_timeout_error(return_value=None)
+@timeout_with_config('WORKFLOWS_DOWNLOAD_DOCUMENT_TIMEOUT')
 def download_documents(obj, eng):
     LOGGER.info('Downloading documents for %s', obj.id)
     documents = obj.data.get('documents', [])
@@ -456,12 +460,12 @@ def download_documents(obj, eng):
         else:
             obj.log.error(
                 'Cannot download document for %s from %s', obj.id, url)
-    delete_empty_key(obj, 'documents')
-    if current_app.config['FEATURE_FLAG_ENABLE_SAVE_WORKFLOW_ON_DOWNLOAD_DOCUMENTS']:
-        save_workflow(obj, eng)
-    else:
-        obj.save()
+    if current_app.config['FEATURE_FLAG_ENABLE_DELETE_EMPTY_KEY_DOCUMENTS']:
+        delete_empty_key(obj, 'documents')
+
+    save_workflow(obj, eng)
     LOGGER.info('Documents downloaded for %s: %s', obj.id, len(obj.data.get('documents', [])))
+    return
 
 
 @backoff.on_exception(backoff.expo, (BadGatewayError, requests.exceptions.ConnectionError), base=4, max_tries=5)