All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=809207

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=3594733bf4c388e6c45d5cf273e5934d3e239de7423a6c5d8ce0e4df10594b4a


Bot(s) for this bug's original alert(s):

chromium-rel-mac-retina

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/16d530d1840000

+ Media Perf Sheriff.

Looks like a BattOr bug:

Traceback (most recent call last):
  File "/b/s/w/ir/third_party/catapult/telemetry/telemetry/internal/platform/tracing_controller_backend.py", line 78, in _CollectNonfatalException
    yield
  File "/b/s/w/ir/third_party/catapult/telemetry/telemetry/internal/platform/tracing_controller_backend.py", line 112, in StartTracing
    if agent.StartAgentTracing(config, timeout):
  File "/b/s/w/ir/third_party/catapult/telemetry/telemetry/internal/platform/tracing_agent/battor_tracing_agent.py", line 73, in StartAgentTracing
    self._battor.StartTracing()
  File "/b/s/w/ir/third_party/catapult/common/battor/battor/battor_wrapper.py", line 234, in StartTracing
    self._SendBattOrCommand(self._START_TRACING_CMD)
  File "/b/s/w/ir/third_party/catapult/common/battor/battor/battor_wrapper.py", line 358, in _SendBattOrCommand
    'Outputted: %s' % (cmd, status))
BattOrError: BattOr did not complete command 'StartTracing' correctly.
Outputted: [0201/010516.085844:FATAL:battor_agent_bin.cc(100)] Fatal error when communicating with the BattOr: TOO MANY COMMAND RETRIES

https://logs.chromium.org/v/?s=chrome%2Fbb%2Fchromium.perf%2FMac_Retina_Perf%2F1746%2F%2B%2Frecipes%2Fsteps%2Fmedia.desktop_on_Intel_GPU_on_Mac_on_Mac-10.11%2F0%2Fstdout

📍 Couldn't reproduce a difference.
https://pinpoint-dot-chromeperf.appspot.com/job/16d530d1840000

This is actually an interaction between two different bugs: 

- https://bugs.chromium.org/p/chromium/issues/detail?id=755981, where BattOr tests time out rather than Telemetry failing if a BattOr error occurs
- https://bugs.chromium.org/p/chromium/issues/detail?id=780907, which greatly increases the likelihood that, especially on Mac, BattOrs will fail when downloading traces and eventually cause the BattOr to become unresponsive

The BattOr on this mac became unresponsive, causing all BattOr-using benchmarks on this host to start to time out instead of running normally. The problem went away when the BattOr was reset (in response to another bug), as you can see by the recent drop in the benchmark duration.

aschulman@ has a pending CL (https://chromium-review.googlesource.com/c/chromium/src/+/908252) which should drastically improve the BattOr resilience.