If everything works fine, but your Apache log keeps showing those errors likely some other systems, perhaps a monitoring system is accessing port 8088 on those backends directly without the proxy protocol.
If it doesn’t work, either component may not have the configuration applied fully or is buggy.
It does work (at least as far as I can tell). The IP address is from the HAProxy host, and there isn’t anything else running on that machine that would hit the Apache backend.
You have posted a correct 5 line snippet of the Apache configuration and a correct 3 line snippet of the haproxy configuration.
Other than saying that this part of the configuration is not wrong, there is really nothing to be done remotely.
Either you provide more information (like the full configuration with private data replaced instead of removed), or your capture the traffic and analyze the proxy protocol traffic.
I’ve narrowed this down to HAProxy health checks. When send-proxy-v2 is combined with check*. *The errors appear every ~2 seconds per server slot, matching the default health check interval. Real proxied traffic works fine.