Reposting from GitHub https://github.com/containous/traefik-helm-chart/pull/56#issuecomment-689853926
This is extremely disappointing to see the Traefik team refusing to support a valid deployment strategy that was used previously to great effect. If you're going make the helm chart highly opinionated then it should only be provided along with the standard YAML as part of the install documentation https://docs.traefik.io/getting-started/install-traefik/.
In my use case I ran Traefik 1.x in DaemonSet because, gasp, there are some people who DON'T run in the cloud. I don't have load balancers in front of my servers. Instead I provide load balancing via DNS RR. I need those hosts to all be running Traefik with 100% guarantee. The correct way to assure one and exactly one copy of a pod is running on every host is a DaemonSet. This is not my opinion, it's in the official docs. Anti-Affinity hacks are not the recommended solution, DaemonSet is. Oh and since I'm aware I can't have LetsEncrypt on multiple nodes, I run CertManager. As it was previously recommended. It's not rocket science.
It used to be easier to get all this running in Traefik 1.x, now 2.x has obscured setups behind helm charts and made it a huge effort in order to even figure out what the correct "default" configuration is, let alone customize it.