Debugging why WPT waterfall shards spuriously time out.

Lower processes from 10 to 5 (to see if it improves stability), and enable mach logging (which
contains timestamps of test status output).

Waterfall: http://go/wpt-waterfall
Design: http://go/wpt-in-chromium-ci

Change-Id: I5f982bf6ea724e6136753bf583f13b438bee2522
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1900167
Commit-Queue: Stephen Martinis <martiniss@chromium.org>
Auto-Submit: Luke Z <lpz@chromium.org>
Reviewed-by: Stephen Martinis <martiniss@chromium.org>
Cr-Commit-Position: refs/heads/master@{#713028}

testing/scripts/run_wpt_tests.py

@@ -54,7 +54,7 @@ class WPTTestAdapter(common.BaseIsolatedScriptArgsAdapter):
             "--headless",
             # TODO(lpz): Consider removing --processes and compute automatically
             # from multiprocessing.cpu_count()
-            "--processes=10",
+            "--processes=5",
             "--no-capture-stdio",
             "--no-manifest-download",
             "--no-pause-after-test",
@@ -71,7 +71,9 @@ class WPTTestAdapter(common.BaseIsolatedScriptArgsAdapter):
             # update the manifest in cast it's stale.
             #"--no-manifest-update",
             "--manifest=../../third_party/blink/web_tests/external/"
-                "WPT_BASE_MANIFEST_6.json"
+                "WPT_BASE_MANIFEST_6.json",
+            # Enable mach logging to aid debugging
+            "--log-mach -"
         ])
         return rest_args