[RLlib] Fix test_actor_manager CI test.

[sven1977]

Fix test_actor_manager CI test.

• Making timeout even more conservative.

The root cause of this test failing is that we are sending out 2x4 ray remote requests to 4 actors (methods "call()" and "ping()") through the ActorManager's async_req utility, which is nothing else other than making the ray remote call on the actors and returning the futures (Ray ObjRefs). Then we sleep, then do a ray.wait() to collect the results.

Any slowdowns here in the past are due to the actors' requests not being scheduled on time due to possibly other actors on the same node taking priority.

Why are these changes needed?

Related issue number

Checks

• I've signed off every commit(by using the -s flag, i.e., git commit -s) in this PR.
• I've run scripts/format.sh to lint the changes in this PR.
• I've included any doc changes needed for https://docs.ray.io/en/master/.
  • I've added any new APIs to the API Reference. For example, if I added a
    method in Tune, I've added it in doc/source/tune/api/ under the
    corresponding .rst file.
• I've made sure the tests are passing. Note that there might be a few flaky tests, see the recent failures at https://flakey-tests.ray.io/
• Testing Strategy
  • Unit tests
  • Release tests
  • This PR is not tested :(

[can-anyscale]

the test_actor_manager, test_bc, etc. doesn't run on microcheck/premerge. because they are marked as flaky tests in this file https://github.com/ray-project/ray/blob/master/ci/ray_ci/rllib.tests.yml#L21. Maybe remove them from the file so that they can run on permerge, thankks

[simonsays1980]

LGTM.