At least on my linux desktop, running the following consistently fails AppToExtension

ninja -C out/Default -j 1000 unit_tests && out/Default/unit_tests --gtest_filter=ExtensionServiceSyncTest.*
...
[ RUN      ] ExtensionServiceSyncTest.AppToExtension
../../chrome/browser/extensions/extension_service_sync_unittest.cc:2608: Failure
      Expected: 1u
      Which is: 1
To be equal to: apps_processor.data().size()
      Which is: 0
../../chrome/browser/extensions/extension_service_sync_unittest.cc:2609: Failure
      Expected: 1u
      Which is: 1
To be equal to: apps_processor.changes().size()
      Which is: 0
[  FAILED  ] ExtensionServiceSyncTest.AppToExtension (107 ms)
[----------] 10 tests from ExtensionServiceSyncTest (5198 ms total)

[----------] Global test environment tear-down
[==========] 10 tests from 1 test case ran. (5199 ms total)
[  PASSED  ] 9 tests.
[  FAILED  ] 1 test, listed below:
[  FAILED  ] ExtensionServiceSyncTest.AppToExtension


Bizarrely, just running AppToExtension by itself, it consistently succeeds.

> ninja -C out/Default -j 1000 unit_tests && out/Default/unit_tests --gtest_filter=ExtensionServiceSyncTest.AppToExtension
...
[1/1] ExtensionServiceSyncTest.AppToExtension (179 ms)
SUCCESS: all tests passed.

It seems if you run all of unit_tests it fails the first time and then retries later and succeeds, thus passing CQ. Not quite sure what state is shared between test runs that is causing this behavior.