Similar to issue 813694 this test and multiple other tests are hitting a NOTIMPLEMENTED.

This test hasn't been touched for a while and is not expected to have timeouts.

According to issue 813694#c2 the NOTIMPLEMENTED will be gone soon and I'd expect the flakiness to go away.
jonross@ Does that sound reasonable or do you think there should be other issues about this test?

Thanks!

```
05:32:09.196 24381 worker/2 virtual/gpu/fast/canvas/canvas-composite.html output stderr lines:
05:32:09.196 24381   
05:32:09.196 24381   DevTools listening on ws://127.0.0.1:37642/devtools/browser/f6df752b-9943-4f99-a272-34218f311269
05:32:09.196 24381   Fontconfig warning: "/etc/fonts/fonts.conf", line 146: blank doesn't take any effect anymore. please remove it from your fonts.conf
05:32:09.196 24381   [1:1:0219/053203.428995:ERROR:single_thread_proxy.cc(648)] Not implemented reached in virtual void cc::SingleThreadProxy::SetRenderFrameObserver(std::unique_ptr<RenderFrameMetadataObserver>)
05:32:09.209 20537 [5/10] virtual/gpu/fast/canvas/canvas-composite.html failed unexpectedly (test timed out)
05:32:09.207 24381 worker/2 killing primary driver
05:32:09.209 24381 worker/2 killing secondary driver
05:32:09.209 24381 worker/2 virtual/gpu/fast/canvas/canvas-composite.html failed:
05:32:09.209 24381 worker/2  test timed out
```

I've submitted a change to get ride of the NOTIMPLEMENTED: https://chromium-review.googlesource.com/c/chromium/src/+/926561

But if a single log is enough to make this test flaky, then I'm quite concerned. As you mentioned, there's no listing of flake expectation for this test.

With the test not touched in years, I don't see a clear owner, or way to progress further. If flaking remains after that CL lands, then we could pursue other steps.

Thanks for the change! Will see if the flakiness goes away.

Linking other similar failures.

Flakes seemed to have stopped after that patch landed. So closing out.