benchmarks: Graduate startup.mobile from experimental

Recently the start_with_url.{cold,warm}.startup_pages became flaky and
were disabled. The unfortunate consequence is that the corresponding
dashboard graphs are going silent. The experimental.mobile.startup
benchmark has already bare minimum functionality to start running on
bots (WPR support, meaningful results for cold startup with intent).
Some data is better than no data at all.

Rename 'experimental.startup.mobile' to 'startup.mobile'. Enabling the
runs on bots will be done separately.

Bug: 760498
Cq-Include-Trybots: master.tryserver.chromium.perf:obbs_fyi
Change-Id: Ia6ffb71a4b20d8969c5bef9febd3c5d8e04a82ff
Reviewed-on: https://chromium-review.googlesource.com/1202665Reviewed-by: Ned Nguyen <nednguyen@google.com>
Commit-Queue: Egor Pasko <pasko@chromium.org>
Cr-Commit-Position: refs/heads/master@{#588421}

tools/perf/benchmarks/startup_mobile.py

@@ -22,26 +22,9 @@ from devil.android.sdk import intent # pylint: disable=import-error
 
 # Chrome Startup Benchmarks for mobile devices (running Android).
 #
-# This set of benchmarks (experimental.startup.mobile) aims to replace
-# the benchmark experimental.startup.android.coldish. It brings two
-# improvements:
-# 1. a name that is more aligned with the end state :)
-# 2. uses the new shared state pattern as recommended by:
-#    third_party/catapult/telemetry/examples/benchmarks/android_go_benchmark.py
-# The shared state allows starting the browser multiple times during a
-# benchmark, properly evicting caches, populating caches, running as many
-# iterations as needed without being affected by pageset-repeat being set by
-# bisect/Pinpoint.
-#
-# Note: this benchmark is not yet ready for FYI bots. When we ensure that
-# evicting OS pagecache happens at proper times, we will replace
-# 'experimental.startup.android.coldish' with 'experimental.startup.mobile'.
-# After adding warm starts the resulting benchmark can replace the
-# 'start_with_url.*' family that has aged quite a bit.
-#
 # The recommended way to run this benchmark is:
 # shell> CHROMIUM_OUTPUT_DIR=gn_android/Release tools/perf/run_benchmark \
-#   -v experimental.startup.mobile --browser=android-chrome \
+#   -v startup.mobile --browser=android-chrome \
 #   --output-dir=/tmp/avoid-polluting-chrome-tree
 
 class _MobileStartupSharedState(story_module.SharedState):
@@ -191,4 +174,4 @@ class MobileStartupBenchmark(perf_benchmark.PerfBenchmark):
 
   @classmethod
   def Name(cls):
-    return 'experimental.startup.mobile'
+    return 'startup.mobile'

tools/perf/core/bot_platforms.py

@@ -9,7 +9,7 @@ from core import benchmark_finders
 
 UNSCHEDULED_TELEMETRY_BENCHMARKS = set([
   'experimental.startup.android.coldish',
-  'experimental.startup.mobile',
+  'startup.mobile',
 ])
 
 
@@ -30,8 +30,7 @@ _ANDROID_GO_BENCHMARK_NAMES = {
     'system_health.memory_mobile',
     'system_health.common_mobile',
     'power.typical_10_mobile',
-    'start_with_url.cold.startup_pages',
-    'start_with_url.warm.startup_pages',
+    'startup.mobile',
     'system_health.webview_startup',
     'v8.browsing_mobile',
     'speedometer',

tools/perf/core/test_data/benchmarks_to_shard.json

@@ -403,7 +403,7 @@
     "stories": [
       "intent:coldish:bbc"
     ], 
-    "name": "experimental.startup.mobile"
+    "name": "startup.mobile"
   }, 
   {
     "repeat": 1,