Now that I've built up enough experience to oversee the impact of migrating to v2, I am wondering how I can do so without major downtime.
In my scenario I have a good 300 endpoints sitting behind Traefik v1, I can't simply bring up a Traefik v2 next to it, as obviously the ports are already bound (80, 443).
So, my current thought is to prepare all instances with the v2 labels, spin up a traefik v2 instance on a separate node, verify through the API/Dashboard that all looks as expected and when it does upgrade the running (replicated) instances.
Any remarks/thoughts/better suggestions?
The other thing I was wondering, in v1 I used the consul backend for ACME. The benefit of this was that several distinct traefik instances spread across different AWS AZ's could access the same LE data. In v2 this seems to be unsupported, due to the aforementioned method supposedly being buggy (never experienced issues myself). What's the recommended approach here? EFS or S3 as a backing store?