Observe these three pages, all related to the same
recent CQ run:

master-paladin:
    https://uberchromegw.corp.google.com/i/chromeos/builders/master-paladin/builds/13065
stumpy-paladin:
    https://luci-milo.appspot.com/buildbot/chromeos/stumpy-paladin/26850
HWTest logs from the failed stumpy paladin:
    https://uberchromegw.corp.google.com/i/chromeos/builders/stumpy-paladin/builds/26850/steps/HWTest%20%5Bbvt-cq%5D/logs/stdio

The stumpy paladin (along with many others) failed HWTest because
the lab was closed.  That was correct, but the reporting makes it
too hard to see:
  * The master paladin reports this message:
        stumpy-paladin: The HWTest [bvt-cq] stage failed: ** HWTest did not complete due to infrastructure issues (code 3) **
  * The stumpy paladin stages report NO MESSAGE AT ALL.
  * You have to click down to the HWTest logs to see the
    (reasonably straightforward) message:
      TestLabException: Chromium OS Test Lab is closed: Lab is closed (scheduled Ganeti maintenance in CBF).
      Will return from run_suite with status: INFRA_FAILURE

It's too much work to find the cause; we should do better.
At minimum, the paladin slave should have included the
same message about "infrastructure issues" as the master.
Better still would be to surface the error message in the test
logs.