A connection pool acquisition timeout error should never result in a connection being closed, since the connection in question was never used for anything. However, it appears this can be the case:

{"t":{"$date":"2025-02-10T22:00:22.956+00:00"},"s":"I",  "c":"CONNPOOL", "id":22566,   "ctx":"conn12","msg":"Ending connection due to bad connection status","attr":{"hostAndPort":"localhost:27017","error":"MaxTimeMSExpired: Remote command timed out while waiting to get a connection from the pool, took 118ms, timeout was set to 106ms","numOpenConns":374}} 

This contributes to needless connection churn. We should ensure that errors that indicate the connection was never used do not result in its closure.

This has been observed in 7.0, though I'm not sure if it affects 8.0 after the recent NetworkInterfaceTL refactor.