Example log:

https://luci-logdog.appspot.com/v/?s=chromium%2Fbb%2Fchromium.gpu%2FMac_Release__Intel_%2F68192%2F%2B%2Frecipes%2Fsteps%2Fmaps_pixel_test_on_Intel_GPU_on_Mac_on_Mac-10.12%2F0%2Fstdout

Got a TimeoutException in Telemetry.

Haven't yet seen any failures after build 68192.

This doesn't appear to be failing very frequently, and it might have been a transient issue.

Not sure what's going on. Both of the failing shards:
https://chromium-swarm.appspot.com/task?id=37548e6dd6139710&refresh=10&show_raw=1
https://chromium-swarm.appspot.com/task?id=375481e88f86a210&refresh=10&show_raw=1

ran on the same piece of hardware, build741-m4:
https://chromium-swarm.appspot.com/bot?id=build741-m4&sort_stats=total%3Adesc

The Raritan KVM connection seems to be OK; the PCI IDs for the device and vendor are reported correctly.

I don't see a pattern of failures on the bot right now that would warrant a deeper investigation of the hardware.

There is an ongoing rewrite of Telemetry's WPR server in  Issue 730036  that may be related to these timeouts. Referencing that bug.

Thanks for the report; please continue to file any flakes you see. Closing this one as WontFix 

rewrite of Telemetry's WPR server is happening but the new implementation is not turned on anywhere on the bots yet
This is unrelated to  Issue 730036 .

Thanks xunjieli -- we're eagerly awaiting the rollout of the new WPR server because I think these flaky timeouts are preexisting bugs in the current Python implementation.

 Issue 747540  has been merged into this issue.