I have been experimenting with traefik on a new k8s cluster and have been really impressed so far. Very simple to set up and easy configuration.
I was looking for resources on RAM and CPU resource limits recommendations for traefik in a production environment. I know it would depend on how much traffic there is, but are there some basic guidelines of what to follow?
You should have a pretty good idea of your requirements based on your usage in your staging environment
If you are unable to use your staging environment as a basis (for whatever reason ) you can deploy with no limits, and monitor for the short term, and then set high limits, and work your way down.
As for memory recommendations, TLS certificates use a lot of memory, routing much less so.
If you only have 10 routes, and a few certificates, 100MB of ram may be more than sufficient.
For CPU usage, TLS connection negotiation is CPU intensive (as it is intended), and general TCP/HTTP connections are much less so.
So you are correct, there are no set rules, as we have users that have Traefik use 10GB of ram, and this is perfectly normal, and we have users that will never use 100MB of ram.
Hope this helps