Upgrade to Matomo 5 brought back internal IPs

Support & Bugs
1

I’m running Matomo in Kubernetes (K3S) with the default Traefik ingress. All the visits are logged as having the container IP (10.42..). I fixed it in Matomo 4 by adding:

proxy_client_headers[] = "HTTP_X_FORWARDED_FOR"
proxy_ips[] = "192.168.*.*/16"
proxy_ips[] = "10.*.*.*/8"
proxy_host_headers[] = "HTTP_X_FORWARDED_HOST"

I also followed a post on upgrading to Matomo 5 and added:

proxy_ip_read_last_in_list = 0

But I’m still stuck with the 10.42.. IPs. Has anyone else solved this?

2

Did you also reapply the other proxy settings (client headers)? It seems, unfortunately, that all proxy settings are reset / lost when upgrading from 4.x to 5.x.

See https://matomo.org/faq/how-to-install/faq_98/

1 Like
3

I did… Didn’t help unfortunately.

4

For debugging, if possible, create a PHP scsript in the Matomo root folder with the content:

<?php

var_dump($_SERVER);

And call it from the “outside”. Then, you’ll see a lot of different attributes of the server request. Probably, you’ll be able to see a hint on what’s missing/wrong (e.g no forwarded-for header, or it has another name, or the proxy IPs are different, etc.)

1 Like
5

Great tip!

All of the IP headers are wrong…

HTTP_X_FORWARDED_FOR
REMOTE_ADDR

So this is not a Matomo issue, it’s a Traefik issue.

Thanks for your help!

1 Like