Started bisect job https://chromeperf.appspot.com/buildbucket_job_status/8986409669230323232

Test: src/tools/perf/run_benchmark -v --browser=android-chromium --output-format=chartjson --upload-results --pageset-repeat=1 --also-run-disabled-tests v8.browsing_mobile_turbo
  Metric: v8_execution_cpu_total_max/v8_execution_cpu_total_max

Hi Annie, hi Simon,
did anything happen with the Nexus 5x bot. It seems that it was off and afterwards one of our metrics significantly regressed. Any ideas?

https://chromeperf.appspot.com/report?sid=a9a5f6c8f0655d89d7cf404ca28d361fb1a499e7b479b0c90d3b5eef62d3bb91

Sorry this is hard to read, we're working on making it more visible in the dashboard, but there is a way to check if the devices under test changed. On the tooltip before and after the jump, you can click "Buildbot status page" and you get to the following pages:
https://build.chromium.org/p/chromium.perf/builders/Android%20Nexus5X%20Perf%20%281%29/builds/4573
https://build.chromium.org/p/chromium.perf/builders/Android%20Nexus5X%20Perf%20%281%29/builds/4574

Search for the "device_status" step and it lists all the connected phones and their serials. You can see the serial numbers are different before/after the jump. So yes it looks like a big device swap.

OK, cool. I will close this bug then. Thanks!