Error while Peeking first byte: connection reset by peer

I have a traefik instance which is NATed via a pfSense instance. The pfsense instance does SSL offloading via HAproxy and forwards requests to traefik.

My traefik logs are full of these lines:

time="2019-09-26T12:41:02Z" level=error msg="Error while Peeking first byte: read tcp 172.19.1.2:80->10.10.10.1:21213: read: connection reset by peer"
time="2019-09-26T12:41:04Z" level=error msg="Error while Peeking first byte: read tcp 172.19.1.2:80->10.10.10.1:21231: read: connection reset by peer"

The 172.19.1.2 is the traefik container which is running on a VM in the NATed pfSense network, 10.10.10.1 is the pfSense itself. I'd like to know what service traefik tries to reach which is resetted then. The port number is increased in every line, sometimes in increments of 10, sometimes in increments of 20.

Why do these lines appear, and what can i do to prevent them?

It's not traefik trying to reach another service it's the other way around. The tcp connection comes from 10.10.10.1 to 172.19.1.2 port 80. But whatever caused the connection bails on it immediately, so traefik does not have a chance to respond.

So to prevent this you need to prevent whatever is calling traefik and dropping the connection from doing that.