Ml2 plugin [1] is spawning a worker process to log warning for agent health check for dead agents [2]. But this worker process is not needed for the drivers (example OVN) which won't spawn any agents. This redundant process may consume memory and cpu like other neutron processes.
It would be better if we can push [1] under a configurable option or have any approach where we can avoid spawning this worker for agent less drivers.
Ml2 plugin [1] is spawning a worker process to log warning for agent health check for dead agents [2]. But this worker process is not needed for the drivers (example OVN) which won't spawn any agents. This redundant process may consume memory and cpu like other neutron processes.
It would be better if we can push [1] under a configurable option or have any approach where we can avoid spawning this worker for agent less drivers.
[1] https:/ /github. com/openstack/ neutron/ blob/master/ neutron/ plugins/ ml2/plugin. py#L242 /github. com/openstack/ neutron/ blob/master/ neutron/ db/agents_ db.py#L319
[2] https:/