diff --git a/taskcluster/taskgraph/transforms/tests.py b/taskcluster/taskgraph/transforms/tests.py
index 5ac6776a717d..b6e297fefe66 100644
--- a/taskcluster/taskgraph/transforms/tests.py
+++ b/taskcluster/taskgraph/transforms/tests.py
@@ -788,7 +788,8 @@ def split_chunks(config, tests):
     them and assigning 'this-chunk' appropriately and updating the treeherder
     symbol."""
     for test in tests:
-        if test['suite'].startswith('test-verify'):
+        if test['suite'].startswith('test-verify') or \
+           test['suite'].startswith('test-coverage'):
             env = config.params.get('try_task_config', {}) or {}
             env = env.get('templates', {}).get('env', {})
             test['chunks'] = perfile_number_of_chunks(env.get('MOZHARNESS_TEST_PATHS', ''),
diff --git a/taskcluster/taskgraph/util/perfile.py b/taskcluster/taskgraph/util/perfile.py
index f5bc2c5d31fa..c65987c305ef 100644
--- a/taskcluster/taskgraph/util/perfile.py
+++ b/taskcluster/taskgraph/util/perfile.py
@@ -22,16 +22,16 @@ def perfile_number_of_chunks(try_task_config, head_repository, head_rev, type):
     # TODO: Make this flexible based on coverage vs verify || test type
     tests_per_chunk = 10.0
 
-    if type.startswith('test-verify-wpt'):
+    if type.startswith('test-verify-wpt') or type.startswith('test-coverage-wpt'):
         file_patterns = ['testing/web-platform/tests/**',
                          'testing/web-platform/mozilla/tests/**']
-    elif type.startswith('test-verify-gpu'):
+    elif type.startswith('test-verify-gpu') or type.startswith('test-coverage-gpu'):
         file_patterns = ['**/*webgl*/**/test_*',
                          '**/dom/canvas/**/test_*',
                          '**/gfx/tests/**/test_*',
                          '**/devtools/canvasdebugger/**/browser_*',
                          '**/reftest*/**']
-    elif type.startswith('test-verify'):
+    elif type.startswith('test-verify') or type.startswith('test-coverage'):
         file_patterns = ['**/test_*',
                          '**/browser_*',
                          '**/crashtest*/**',
@@ -69,7 +69,7 @@ def perfile_number_of_chunks(try_task_config, head_repository, head_rev, type):
 
             if mozpackmatch(path, pattern):
                 gpu = False
-                if type == 'test-verify-e10s':
+                if type == 'test-verify-e10s' or type == 'test-coverage-e10s':
                     # file_patterns for test-verify will pick up some gpu tests, lets ignore
                     # in the case of reftest, we will not have any in the regular case
                     gpu_dirs = ['dom/canvas', 'gfx/tests', 'devtools/canvasdebugger', 'webgl']
diff --git a/testing/mozbase/mozlog/mozlog/handlers/statushandler.py b/testing/mozbase/mozlog/mozlog/handlers/statushandler.py
index d145f73874e2..c6c3adac542f 100644
--- a/testing/mozbase/mozlog/mozlog/handlers/statushandler.py
+++ b/testing/mozbase/mozlog/mozlog/handlers/statushandler.py
@@ -30,12 +30,16 @@ class StatusHandler(object):
         self.action_counts = defaultdict(int)
         # The count of messages logged at each log level
         self.log_level_counts = defaultdict(int)
+        # The count of "No tests run" error messages seen
+        self.no_tests_run_count = 0
 
     def __call__(self, data):
         action = data['action']
         self.action_counts[action] += 1
 
         if action == 'log':
+            if data['level'] == 'ERROR' and data['message'] == 'No tests ran':
+                self.no_tests_run_count += 1
             self.log_level_counts[data['level']] += 1
 
         if action in ('test_status', 'test_end'):
diff --git a/testing/mozharness/mozharness/mozilla/structuredlog.py b/testing/mozharness/mozharness/mozilla/structuredlog.py
index 98654acff118..4adaeda4a43e 100644
--- a/testing/mozharness/mozharness/mozilla/structuredlog.py
+++ b/testing/mozharness/mozharness/mozilla/structuredlog.py
@@ -68,7 +68,7 @@ class StructuredOutputParser(OutputParser):
         try:
             candidate_data = json.loads(line)
             if (isinstance(candidate_data, dict) and
-                'action' in candidate_data and candidate_data['action'] in self.log_actions):
+               'action' in candidate_data and candidate_data['action'] in self.log_actions):
                 data = candidate_data
         except ValueError:
             pass
@@ -124,27 +124,37 @@ class StructuredOutputParser(OutputParser):
             1) Remove previous data from the new summary to only look at new data
             2) Build a joined summary to include the previous + new data
         """
+        RunSummary = namedtuple("RunSummary",
+                                ("unexpected_statuses",
+                                 "expected_statuses",
+                                 "log_level_counts",
+                                 "action_counts"))
+        if previous_summary == {}:
+            previous_summary = RunSummary(defaultdict(int),
+                                          defaultdict(int),
+                                          defaultdict(int),
+                                          defaultdict(int))
         if previous_summary:
-            RunSummary = namedtuple("RunSummary",
-                                    ("unexpected_statuses",
-                                     "expected_statuses",
-                                     "log_level_counts",
-                                     "action_counts"))
-
             self.tbpl_status = TBPL_SUCCESS
-
             joined_summary = summary
 
             # Remove previously known status messages
-            summary = RunSummary(self._subtract_tuples(previous_summary.unexpected_statuses, summary.unexpected_statuses),
-                                 self._subtract_tuples(previous_summary.expected_statuses, summary.expected_statuses),
-                                 summary.log_level_counts,
+            if 'ERROR' in summary.log_level_counts:
+                summary.log_level_counts['ERROR'] -= self.handler.no_tests_run_count
+
+            summary = RunSummary(self._subtract_tuples(previous_summary.unexpected_statuses,
+                                                       summary.unexpected_statuses),
+                                 self._subtract_tuples(previous_summary.expected_statuses,
+                                                       summary.expected_statuses),
+                                 self._subtract_tuples(previous_summary.log_level_counts,
+                                                       summary.log_level_counts),
                                  summary.action_counts)
 
-            # If we have previous data to ignore, cache it so we don't parse the log multiple times
+            # If we have previous data to ignore,
+            # cache it so we don't parse the log multiple times
             self.summary = summary
         else:
-           joined_summary = summary
+            joined_summary = summary
 
         fail_pair = TBPL_WARNING, WARNING
         error_pair = TBPL_FAILURE, ERROR
diff --git a/testing/mozharness/scripts/android_emulator_unittest.py b/testing/mozharness/scripts/android_emulator_unittest.py
index 118a39a7d26b..858c5de6d9e9 100644
--- a/testing/mozharness/scripts/android_emulator_unittest.py
+++ b/testing/mozharness/scripts/android_emulator_unittest.py
@@ -774,7 +774,7 @@ class AndroidEmulatorTest(TestingMixin, BaseScript, MozbaseMixin, CodeCoverageMi
             env['MINIDUMP_SAVE_PATH'] = self.query_abs_dirs()['abs_blob_upload_dir']
             env['RUST_BACKTRACE'] = 'full'
 
-            summary = None
+            summary = {}
             for per_test_args in self.query_args(per_test_suite):
                 if (datetime.datetime.now() - self.start_time) > max_per_test_time:
                     # Running tests has run out of time. That is okay! Stop running
diff --git a/testing/mozharness/scripts/desktop_unittest.py b/testing/mozharness/scripts/desktop_unittest.py
index 67a278332476..10a922bb53cc 100755
--- a/testing/mozharness/scripts/desktop_unittest.py
+++ b/testing/mozharness/scripts/desktop_unittest.py
@@ -848,7 +848,7 @@ class DesktopUnittest(TestingMixin, MercurialScript, MozbaseMixin,
                 env = self.query_env(partial_env=env, log_level=INFO)
                 cmd_timeout = self.get_timeout_for_category(suite_category)
 
-                summary = None
+                summary = {}
                 executed_too_many_tests = False
                 for per_test_args in self.query_args(suite):
                     # Make sure baseline code coverage tests are never
diff --git a/testing/mozharness/scripts/web_platform_tests.py b/testing/mozharness/scripts/web_platform_tests.py
index 7f3cd409972f..35469ab71df3 100755
--- a/testing/mozharness/scripts/web_platform_tests.py
+++ b/testing/mozharness/scripts/web_platform_tests.py
@@ -335,7 +335,7 @@ class WebPlatformTest(TestingMixin, MercurialScript, CodeCoverageMixin):
             if suite:
                 test_types = [suite]
 
-            summary = None
+            summary = {}
             executed_too_many_tests = False
             for per_test_args in self.query_args(suite):
                 # Make sure baseline code coverage tests are never