DRIVERS-3134: pre-populate pool in flaky CSOT runCommandCursor test

[baileympearson]

The "Non-tailable cursor lifetime remaining timeoutMS applied to getMore if timeoutMode is unset" flakes fairly regularly in Node with the following error:

 AssertionError: Expected event count mismatch, expected [ 'commandStartedEvent', 'commandStartedEvent' ] but got [ 'CommandStartedEvent' ]

Turns out connection establishment is taking roughly 40ms worth of time. This seems unusually high, measuring connection establishment while running this test repeatedly usually shows establishment taking 2-8 ms. I also observe that, as the tests run, we occasionally hit periods where establishment increases to between 20-40ms for a few iterations of the tests, and then drops back down to the usual 2-8ms for establishment.

The delay in connection establishment causes that the find to time out, not the getMore. As a result, we still get a timeout error in the test but we have too few CommandStartedEvents and the test fails.

Node has seen this in other non-spec tests we wrote while implementing CSOT. The most reliable way to resolve the flakiness is to pre-populate the pool with a connection before starting the test. For any test that relies on failpoints to cause timeouts, it makes sense to pre-populate the pool with a connection so that we can as much variance in timing before the connection layer as possible.

Please complete the following before merging:

• Update changelog.
• Test changes in at least one language driver.
• Test these changes against all server versions and topologies (including standalone, replica set, sharded
  clusters, and serverless).