Mongod instance report a high number of "awaitingTopologyChanges" connections


I’m running a PSA replica set with mongod instances v4.4.5. The Primary and Arbiter instances are on the same server, all running on Ubuntu 20.04.4 .

It’s come to my attention that when I check the connections from server status on both the Primary and Secondary instances, the value of connections awaiting topology changes is disproportionally large.

    current: 49,
    available: 51151,
    totalCreated: 407521,
    active: 16,
    exhaustIsMaster: 3,
    exhaustHello: 12,
    awaitingTopologyChanges: 203193

I understood that this number reflects

The number of clients currently waiting in a hello or isMaster request for a topology change.

I have quite a few questions about this.

  1. Is this an expected status under heavy loads?
  2. What are the implications of this high number on the server performance and network usage?
  3. Why and when do clients wait in a isMaster or Hello requests?
  4. What would be considered a standard procedure in this case, to fix the issue and reduce the possibility of this happening in the future?