Migrate cluster telemetry benchmarks to tools/perf/contrib/cluster_telemetry/

tools/perf/benchmarks/benchmark_smoke_unittest.py

Index: tools/perf/benchmarks/benchmark_smoke_unittest.py
diff --git a/tools/perf/benchmarks/benchmark_smoke_unittest.py b/tools/perf/benchmarks/benchmark_smoke_unittest.py
index 5beb23bdf2b51ca667367ecb8e6b247e5f040ec4..2460783e71154b2a69b71024a1589a5ecc887d29 100644
--- a/tools/perf/benchmarks/benchmark_smoke_unittest.py
+++ b/tools/perf/benchmarks/benchmark_smoke_unittest.py
@@ -27,7 +27,6 @@ from benchmarks import jetstream
 from benchmarks import kraken
 from benchmarks import octane
 from benchmarks import rasterize_and_record_micro
-from benchmarks import repaint
 from benchmarks import speedometer
 from benchmarks import v8_browsing
 
@@ -93,7 +92,6 @@ _BLACK_LIST_TEST_MODULES = {
     indexeddb_perf,  # Always fails on Win7 & Android Tests builder.
     octane,  # Often fails & take long time to timeout on cq bot.
     rasterize_and_record_micro,  # Always fails on cq bot.
-    repaint,  # Often fails & takes long time to timeout on cq bot.

[wkorman, 2017/05/12 17:44:23]

This seems worth noting in the CL description. If we fix the flakiness bug
http://crbug.com/715962 would we want to restore the benchmark to run as a smoke
test to catch regression?

[nednguyen, 2017/05/12 20:17:03]

Oh, this is smoke test of repaint benchmarks that are run on perf waterfall,
which is no longer relevant after your removal of them in
https://codereview.chromium.org/2874043003/

     speedometer,  # Takes 101 seconds.
     jetstream,  # Take 206 seconds.
     kraken,  # Flaky on Android, crbug.com/626174.