Ingresses stop working after some time

I have a strange problem in one of my Traefik deployments. It's installed on a RKE2 cluster (I have a few of them, all configured the same way) but this one's giving me headaches.
This cluster has only one ingress. When deployed the first time, it all worked correctly. But we recently realized that the ingress was no longer working. The page was coming back with:

"kind": "Status",
"apiVersion": "v1",
"metadata": {

"status": "Failure",
"message": "Unauthorized",
"reason": "Unauthorized",
"code": 401

I found out that restarting the traefik pod made the ingress working again.... for some time, then failed again. The ingress doesn't provide a cert, as I didn't configured one. The cert comes from the traefik self-signed one. So, on the first occurence, the browser obviously ask to validate the cert, which I'm doing. And subsequent requests works without problems. Eventually, when the ingress starts failing, I'm asked again to validate the self-signed cert, as if that was a new one, but leading to ths error above.
Any clue? It's traefik 2.7.1 (I know, it's old, but working as is on all other clusters), installed from helm with chart 10.23.0