Whatever the problem is, it appears to be specific to the
"cr50_stress_experimental" test suite, not to anything relating to
the M69 milestone.  I see plenty of jobs that suite_scheduler
scheduled against M69 builds, including specifically against
fizz-release/R69-10727.0.0.  However, I can't find any jobs scheduled
against that cr50 suite.

I checked, and there are plenty of instances of "cr50_stress_experimental"
being scheduled against M68, but none against M69.

There's a very similar symptom reported against the wifi tests.

Hmmm...  I note this:

$ atest host list -b pool:cr50_stress_experimental | count_labels -m
      1 bob
      1 electro
      1 eve
      1 robo360
      1 scarlet
      1 soraka
      1 teemo

That will explain lack of testing on fizz.  There's also this:

$ dut-status -p cr50_stress_experimental -m scarlet
hostname                       S   last checked         URL
chromeos1-row2-rack1-host2     NO  2018-05-29 12:16:29  http://cautotest.corp.google.com/tko/retrieve_logs.cgi?job=/results/hosts/chromeos1-row2-rack1-host2/63432422-repair/
: jrbarnette .../src/third_party/autotest/files; dut-status -p cr50_stress_experimental
hostname                       S   last checked         URL
chromeos1-row1-rack5-host2     ??  2018-05-27 10:40:53  http://cautotest.corp.google.com/tko/retrieve_logs.cgi?job=/results/hosts/chromeos1-row1-rack5-host2/1147405-reset/
chromeos1-row1-rack5-host3     NO  2018-05-29 12:38:13  http://cautotest.corp.google.com/tko/retrieve_logs.cgi?job=/results/hosts/chromeos1-row1-rack5-host3/782555-repair/
chromeos1-row1-rack5-host1     NO  2018-05-29 12:07:42  http://cautotest.corp.google.com/tko/retrieve_logs.cgi?job=/results/hosts/chromeos1-row1-rack5-host1/926816-repair/
chromeos1-row2-rack1-host4     NO  2018-05-29 06:57:17  http://cautotest.corp.google.com/tko/retrieve_logs.cgi?job=/results/hosts/chromeos1-row2-rack1-host4/63430473-repair/
chromeos1-row2-rack1-host3     --  ---                  ---
chromeos1-row2-rack1-host2     NO  2018-05-29 12:16:29  http://cautotest.corp.google.com/tko/retrieve_logs.cgi?job=/results/hosts/chromeos1-row2-rack1-host2/63432422-repair/
chromeos1-row2-rack1-host5     ??  2018-05-25 15:40:07  http://cautotest.corp.google.com/tko/retrieve_logs.cgi?job=/results/hosts/chromeos1-row2-rack1-host5/1281027-repair/

There's a number of DUTs currently down, which would explain lots of
other lack of testing.

There's also an update on  bug 847540 , which is enough to
explain any remaining lack of testing.  But, even if we
fix how ToT is calculated, the DUT availability will still prevent
testing the cr50 suite at issue here.

To the extent that there's still a problem that _isn't_  bug 847540 ,
the problem is DUTs in the chromeos1 lab.  That's outside the scope
of CrOS Infra.

So...  Back to the OP, who can close this or act on it, as necessary.

thanks for all the investigation

the M69 layer of problem is now well understood

the physical layer of problems are mostly well understood and tracked in b/80306939

the mismatch between model/board noted in c4 I think may need some more attention: will take a closer look once the physical layer problems are cleared (the scheduler syntax is "board", so teemo matching for fizz would be intuitive).  Considering that out of scope.