Are you pushing your agents out? That's a very archaic way of doing it & will fail due to firewalls, security restrictions (things like RPC not being enabled), and so on.
Use the advanced agent instead & push out via login script or many other methods. It'll also automatically make use of peer download & so on, so as not to stress the WAN, while agent push is a unicast every single time.
You'll get a much clearer WSCFG32.XLG log as well - while the 20+ year old "push" method is really just a fallback option.
If you end up needing more help / support here - there's a bunch of BKM's on how to use the advanced agent & so on.
If you get stuck, let us know .