Findit identified the culprit r582117 with confidence 70.0% in the config "chromium.linux / Linux Tests (dbg)(1)"
based on the flakiness trend:

https://findit-for-me.appspot.com/waterfall/flake?key=ag9zfmZpbmRpdC1mb3ItbWVyvQELEhdNYXN0ZXJGbGFrZUFuYWx5c2lzUm9vdCKGAWNocm9taXVtLmxpbnV4L0xpbnV4IFRlc3RzIChkYmcpKDEpLzczNjg2L3dlYmtpdF9sYXlvdXRfdGVzdHMvZG1seWRIVmhiQzkwYUhKbFlXUmxaQzltWVhOMEwybGtiR1V0WTJGc2JHSmhZMnN2YVdSc1pWOXdaWEpwYjJSekxtaDBiV3c9DAsSE01hc3RlckZsYWtlQW5hbHlzaXMYAQw

If the culprit above is wrong, please file a bug using this link and hit submit:
https://bugs.chromium.org/p/chromium/issues/entry?status=Unconfirmed&labels=Pri-1,Test-Findit-Wrong&components=Tools%3ETest%3EFindit%3EFlakiness&summary=%5BFindit%5D%20Flake%20Analyzer%20-%20Wrong%20result%20for%20virtual/threaded/fast/idle-callback/idle_periods.html&comment=Link%20to%20Analysis%3A%20https://findit-for-me.appspot.com/waterfall/flake?key=ag9zfmZpbmRpdC1mb3ItbWVyvQELEhdNYXN0ZXJGbGFrZUFuYWx5c2lzUm9vdCKGAWNocm9taXVtLmxpbnV4L0xpbnV4IFRlc3RzIChkYmcpKDEpLzczNjg2L3dlYmtpdF9sYXlvdXRfdGVzdHMvZG1seWRIVmhiQzkwYUhKbFlXUmxaQzltWVhOMEwybGtiR1V0WTJGc2JHSmhZMnN2YVdSc1pWOXdaWEpwYjJSekxtaDBiV3c9DAsSE01hc3RlckZsYWtlQW5hbHlzaXMYAQw

Automatically posted by the findit-for-me app (https://goo.gl/Ot9f7N).

The findit tasks (see e.g. https://chromium-swarm.appspot.com/task?id=3f3d831086999c10&refresh=10&show_raw=1 ) show error output like:

[1:1:0810/092643.081764:ERROR:runtime_enabled_features.cc(1540)] RuntimeEnabledFeature not recognized: FeaturePolicyForPermissions

It looks like this flag was removed in https://chromium-review.googlesource.com/c/chromium/src/+/1166619 .  I'm wondering if somehow there's some kind of stale reference in a config file, or a bot is out of date, or something.

->owner of that CL to track down further.

Findit blames "Oilpan: Enable incremental marking per default"

The log output
"RuntimeEnabledFeature not recognized: FeaturePolicyForPermissions"
is unlikely to be relevant. It is being fixed in
https://chromium-review.googlesource.com/c/chromium/src/+/1176882

+rmcilroy: Looks like the test does not get enough idle time to pass which is kind of expected if incremental marking takes up this idle time.

Keishi, can you have a look?

The theory is that we do drain all the idle time when a GC kicks in. It would be good to figure out what exactly happens.

Long term, I think we should move the incremental steps out of the safe point trigger and provide them with their own task. This probably means that we need a "GC task queue" to avoid being delayed too much.