diff --git a/run_tests.sh b/run_tests.sh
index 76be1823e68..80fadd5cb68 100755
--- a/run_tests.sh
+++ b/run_tests.sh
@@ -35,6 +35,14 @@ Run a specific API test:
     ./run_tests.sh -api test/api/test_tools.py:ToolsTestCase.test_map_over_with_output_format_actions
 
 
+External Tests:
+
+A small subset of tests can be run against an existing Galxy
+instance. The external Galaxy instance URL can be configured with
+--external_url. If this is set eithre --external_master_key or
+--external_user_key must be set as well - more tests can be executed
+with --external_master_key than with a user key.
+
 Extra options:
  --verbose_errors      Force some tests produce more verbose error reporting.
  --no_cleanup          Do not delete temp files for Python functional tests (-toolshed, -framework, etc...)
@@ -45,6 +53,10 @@ Extra options:
  --dockerize           Run tests in a pre-configured Docker container (must be first argument if present).
  --db <type>           For use with --dockerize, run tests using partially migrated 'postgres', 'mysql',
                        or 'sqlite' databases.
+  --external_url       External URL to use for Galaxy testing (only certain tests).
+  --external_master_key Master API key used to configure external tests.
+  --external_user_key  User API used for external tests - not required if
+                       external_master_key is specified.
 
 Environment Variables:
 
@@ -200,6 +212,18 @@ do
           with_framework_test_tools_arg="-with_framework_test_tools"
           shift
           ;;
+      --external_url)
+          GALAXY_TEST_EXTERNAL=$2
+          shift 2
+          ;;
+      --external_master_key)
+          GALAXY_CONFIG_MASTER_KEY=$2
+          shift 2
+          ;;
+      --external_user_key)
+          GALAXY_TEST_USER_API_KEY=$2
+          shift 2
+          ;;
       -w|-workflow|--workflow)
           if [ $# -gt 1 ]; then
               workflow_file=$2
diff --git a/scripts/functional_tests.py b/scripts/functional_tests.py
index 668f3669ae4..db49adfcda8 100644
--- a/scripts/functional_tests.py
+++ b/scripts/functional_tests.py
@@ -38,14 +38,14 @@ def main():
         default_tool_conf = driver_util.FRAMEWORK_SAMPLE_TOOLS_CONF
         datatypes_conf_override = driver_util.FRAMEWORK_DATATYPES_CONF
 
-    start_server = 'GALAXY_TEST_EXTERNAL' not in os.environ
+    external_galaxy = os.environ.get('GALAXY_TEST_EXTERNAL', None)
 
     galaxy_test_tmp_dir = driver_util.get_galaxy_test_tmp_dir()
 
     app = None
     server_wrapper = None
 
-    if start_server:
+    if external_galaxy is None:
         tempdir = tempfile.mkdtemp( dir=galaxy_test_tmp_dir )
         # Configure the database path.
         galaxy_db_path = driver_util.database_files_path(tempdir)
@@ -65,6 +65,8 @@ def main():
             galaxy_config,
         )
         log.info("Functional tests will be run against %s:%s" % (server_wrapper.host, server_wrapper.port))
+    else:
+        log.info("Functional tests will be run against %s" % external_galaxy)
 
     # ---- Find tests ---------------------------------------------------------
     success = False
@@ -90,19 +92,20 @@ def main():
                 user_api_key=get_user_api_key(),
             )
 
-        # We must make sure that functional.test_toolbox is always imported after
-        # database_contexts.galaxy_content is set (which occurs in this method above).
-        # If functional.test_toolbox is imported before database_contexts.galaxy_content
-        # is set, sa_session will be None in all methods that use it.
-        import functional.test_toolbox
-        functional.test_toolbox.toolbox = app.toolbox
-        # When testing data managers, do not test toolbox.
-        functional.test_toolbox.build_tests(
-            app=app,
-            testing_shed_tools=testing_shed_tools,
-            master_api_key=get_master_api_key(),
-            user_api_key=get_user_api_key(),
-        )
+        if app is not None:
+            # We must make sure that functional.test_toolbox is always imported after
+            # database_contexts.galaxy_content is set (which occurs in this method above).
+            # If functional.test_toolbox is imported before database_contexts.galaxy_content
+            # is set, sa_session will be None in all methods that use it.
+            import functional.test_toolbox
+            functional.test_toolbox.toolbox = app.toolbox
+            # When testing data managers, do not test toolbox.
+            functional.test_toolbox.build_tests(
+                app=app,
+                testing_shed_tools=testing_shed_tools,
+                master_api_key=get_master_api_key(),
+                user_api_key=get_user_api_key(),
+            )
         success = driver_util.nose_config_and_run()
     except:
         log.exception( "Failure running tests" )
diff --git a/test/base/twilltestcase.py b/test/base/twilltestcase.py
index 0a9af77d3c6..3fe965d1835 100644
--- a/test/base/twilltestcase.py
+++ b/test/base/twilltestcase.py
@@ -48,7 +48,8 @@ class TwillTestCase( unittest.TestCase ):
         self.history_id = os.environ.get( 'GALAXY_TEST_HISTORY_ID', None )
         self.host = os.environ.get( 'GALAXY_TEST_HOST' )
         self.port = os.environ.get( 'GALAXY_TEST_PORT' )
-        self.url = "http://%s:%s" % ( self.host, self.port )
+        default_url = "http://%s:%s" % (self.host, self.port)
+        self.url = os.environ.get('GALAXY_TEST_EXTERNAL', default_url)
         self.test_data_resolver = TestDataResolver( )
         self.tool_shed_test_file = os.environ.get( 'GALAXY_TOOL_SHED_TEST_FILE', None )
         if self.tool_shed_test_file: