Simple HTTPs for Kubernetes // Traefik + Cert-Manager

[@everythingiscoolio]

Man, thank you so much. This came at a perfect time for me. The past 3 weeks I've been banging my head against the wall, and I finally just finished setting up both nginx and traefik in my cluster. Great little study project!!! Your video was a great help to me, and I just wanted to share my gratitude! Thank you for doing this, for free!!

[@christianlempa]

Thank you so much, I'm glad it was helpful :)

[@jmulaaclips6859]

For people who aren't using a k8s distro that has kube-vip installed, consider using MetalLB to handle load balancing. Great video!

[@christianlempa]

Thanks! Btw MetalLB is great, I used it on my local test cluster, I just thought why install 2 apps when 1 app does both tasks ;)

[@robfahey7363]

Great video Christian, I ran into Cert-Manager when playing with Rancher. I was amazed I got it running with my internal CA, and traefik with a single wildcard cert. Keep up the good work. I didn't know that some helm deployments can take care of the ingress, that's so much easier.

[@christianlempa]

Awesome! Thank you so much :)

[@freshgrease]

Man it’s been a trip since I followed the original video. Now I’m running a 3 node master on the Pis and supplementing with VMs. I’d also say integrate external-dns if you’re doing traefik and cert manager

[@christianlempa]

Wow, that is amazing! :)

[@SnorreSelmer]

Fantastic! Thanks so much for your really helpful tutorials!

[@christianlempa]

Glad you like them!

[@KaydotOrigin]

👍 this is a great video idea! Appreciate the content

[@christianlempa]

Glad you think so!

[@wstrater]

I did see the link to your Twingate video in your description.

[@christianlempa]

Good point! Added it to the list :D

[@dean.kannenberg]

Great work! Thank you for that great video again :)

[@christianlempa]

Thank you so much 😊

[@DDD-d1g]

What is that interesting shell you are using?

[@DDD-d1g]

BTW it will be interesting to compare your shell customization with that of @DevOpsToolkit and others.

[@christianlempa]

I'm using Warp Terminal

[@fherbet]

Thanks for this great video ! i completed the whole setup and it works perfectly fine ! now i'm stuck with most of helm charts which use nginx ingress and not traefik ingressroute... can you please point me into the right directions to be able to modify helm charts with ingressroute? are there any best practice or detailed guidance for this? thanks again for your awesome videos !

[@christianlempa]

Thanks for the kind words! :) I usually use the Ingress objects when Helm supports it, or I add ingressroute.yaml manifests into my deployment CICD, or I manually apply them later after the Helm deployments.

[@crazywolf13]

Awesome!! But what about the video about gitea/forgejo comparison to gitlab?

[@christianlempa]

Thx! Gitea videos are already planned for December 😊👍

[@crazywolf13]

@christianlempa Awesome 😎

[@lucEast]

One thing I didn't fully understand: Don't you need to expose your domain to the public so that Let's Encrypt can verify the domain?

[@XHACKER3900]

It is with the HTTP challenge. Alternatively you can use DNS challenge to avoid exposing services. While DNS challenge public DNS records have to be set

[@joshuaboley4135]

Love the video! Have a question for you though: Say I have multiple subnets on my cluster (management, public facing, etc.). What sort of impact would that have on the traefik service deployment? It seems like it should be as simple as providing IPs in the subnet I want to use, but in my experience there’s usually a “gotcha” lurking somewhere. Besides, I’m still very new to Kubernetes, so I rather expect to things to spontaneously break while the computer gremlins laugh at the noob 😅

[@christianlempa]

Hey! Thanks for the love! 😊 When it comes to multiple subnets with Traefik, you’re right that it should be straightforward, but there can definitely be some quirks. Make sure your network policies and firewall rules allow traffic between those subnets. Also, check if your load balancer needs specific settings for each subnet. Don't worry too much about the gremlins; everyone starts somewhere, and you’ll get the hang of it! If you hit any snags, just ask!

[@alexdin1565]

please can you make videos about coolify?

[@christianlempa]

Hm I don't know what I should do with it tbh

[@rafal9ck817]

if you use echo | base64 remember -n flag or else you will get error in cert events

[@christianlempa]

Good point! ;)

[@b14ckh4wk3]

I bet you don't even get 100req/s and you are using k8s? I don't understand this generation of bad DevOps guys, they want to resolve everything with k8s, docker among other owned and unnecessary abstractions

[@pesfreak18]

Is it useful in a homelab? Probably not. But that applys to almost all service in a homelab. Still it's fun to tinker around with these tools and maybe the knowledge you gain will be useful someday. If not than at least you can have fun with a hobby.

[@DDD-d1g]

It's a good play ground to learn k8s in order to use it in your real job in a production environment of a big enterprise.

[@EpicGamingNL]

Thats the best way to learn; you can break anything and not many people will be affected by it, well maybe friends and family but those are my test users lol, the experience can later be used in production environments and can help with your job :)

[@everythingiscoolio]

Reasons to adopt k8s do include high load, but it is far from the only reason. Calling other people "bad" while you don't understand literally the most basic shit any graduate has been able to teach themselves is pretty funny. Maybe you should stop giving your opinion about things you know nothing about?

[@rafal9ck817]

I have 2 nodes,one can't keep up and I do not want to manualy manage 2 docker vms, I could probably do docker swarm but that'd be too easy