Why does everyone show the same diagram with a load balancer with many arrows coming out of it, one of which is going into a kubernetes cluster? This only serves to confuse people, as if the data from the load balancer is a straight line pipe into a single cluster. This is not how it actually works. The external load balancer talks directly to the nodes themselves. If you're trying to set up a service mesh, maybe this diagram is accurate, but you're simply confusing people if this is not the case.
Great Explanation! But I’m still not sure what to use for my setup, I want to setup a kubernetes cluster in the cloud so each node one VM. My cloud provider has a Load Balancer service but limits the possible targets it can connect to, so I thought I will connect the Cloud Load Balancer to a Ingress/Load Balancer on the Cluster which then routes the traffic accordingly. That way I only use 3 Targets of my load balancer and my Cluster also controls the certs, so I don’t need to pay extra for more certs and load balancer recourses. How could I accomplish that, and is this even possible when wanting a production grade/ready cluster?
You can use the Nginx ingress controller. Set it up using a single 'target,' possibly two for HTTP/S, and then internally route to your services using the Nginx controller.
@@AntonPutra Thanks! So my way wasn’t completely wrong. But as I understand this would now be a single point of failure, so how can I have HA Ingress Controller? Then the external load balancer can point to the 3 or more ingress controller.
Use ingress when you want to share one load balancer among multiple services, which is limited to HTTP/S only. Yes, most of the time you would use ClusterIP, except in some edge cases when you use the AWS Load Balancer Controller in instance mode.
почему так сложно обяснить получше про LoadBalancer service? Ты говоришь, что он открывает под во вне, но совершенно непонятно, нужно навешивается реально какой-то балансировщик нагрузки или нет? например на 5:00 по смыслу да. а в остальном видео по смыслу нет.
I didn't understand the question. In the cloud, the LoadBalancer service creates a cloud load balancer and routes traffic to your app; on-premise, it depends on your configuration. Also, LoadBalancer can use node ports underneath, and if Kubernetes supports native networks (without virtual network plugins such as Flannel, Calico), it can route directly to your pod's IP address. (I don't have a Russian keyboard.)
Well that's how cloud providers used to do it behind the load balancer type service. Nowadays, they utilize IP mode and route traffic directly to the pod. On-premise, perhaps...
➜ ~ kubectl -n kubernetes-dashboard port-forward svc/kubernetes-dashboard 8443:443 Error from server (NotFound): namespaces "kubernetes-dashboard" not found
🔴 - To support my channel, I’d like to offer Mentorship/On-the-Job Support/Consulting - me@antonputra.com
I highly appreciate your effort to teach advanced stuff with such depth and graphics.
Keep them coming
Thanks, will do!
First time understood the ingress concept. Highly Appreciated and Helpful.
Once again, Anton, you are a legend, this is precisely what I wanted to know. Amazing!!! 🔥
Thank you!
Learning k8s right now and this is video is really interesting. Thank you, Anton!
Thanks! Appreciate it! One more k8s video coming in couple of days
I really appreciate your content, thanks Anton!
Thank you!
Exactly what I was looking for 🙏
Thank you, Anton! Really great stuff, one day I will be as knowledgable as you are
Thank you!
Super Helpful Anton..It helped me crack an interview!
glad that it helped!
Great visualization. Thanks for sharing.
Thank you!
Thank you for another great video, Anton. It is very informative.
Thank you! Appreciate it!
This was really good. Thank you. Although the first part of how ClusterIP was exposed outside for the Kubernetes Dashboard kinda went over my head 😅
thanks :) let me know if you have any questions
Looking forward on Kubernetes 1.27: Avoid Collisions Assigning Ports to NodePort Services
What's your use case for NodePorts?
@@AntonPutra chatbot platform with different market region.
@@zekeluther I mean your use case for nodeports, why not to use load balancer or ingress?
dude your content is great, I am watching one after one.. thank you
Thank you!
спасибо ! полезная инфа !
spasibo za otziv!
Your content is pure gold!!
Thank you!
Amazing knowledge
Thanks!
Thank you! Very clear!
thanks!
Very good tutorial with so much details !
thanks!
Thanks for content Anton!
May I ask your opinion, is it worth getting a certificate in Kubernetes(CKA or CKAD)?
Personally, I never pay attention to whether a candidate has certificates or not. Perhaps someone else does.
That's is a perfect answer
Thanks. Very clear and concise
thanks!
Great stuff!
thank you!
Excelente video 🤌🏽🙌🏽
Thank you!
perfect explanation, keep it up!
thanks! will do
Informative content .. Which tool do you use for graphics ?
thanks, adobe suite
nicely explained. Thankyou!
Thanks!
Wow!! Thank you!
Thank you!
I thought you were Indonesian 😅 Putra is a really common name here. Great content btw!
Thanks! 😃 coming back to Bali soon :))
Why does everyone show the same diagram with a load balancer with many arrows coming out of it, one of which is going into a kubernetes cluster? This only serves to confuse people, as if the data from the load balancer is a straight line pipe into a single cluster. This is not how it actually works. The external load balancer talks directly to the nodes themselves. If you're trying to set up a service mesh, maybe this diagram is accurate, but you're simply confusing people if this is not the case.
What do you mean? The arrows on one end are either the ingresses or services themselves and on the other end it's control plane
Very well explained!
Thank you!
Great video. Thank you.
my pleasure!
What a vid. Thanks again
Thanks Rafael!
Very well explained.
thanks!
Great Explanation! But I’m still not sure what to use for my setup, I want to setup a kubernetes cluster in the cloud so each node one VM. My cloud provider has a Load Balancer service but limits the possible targets it can connect to, so I thought I will connect the Cloud Load Balancer to a Ingress/Load Balancer on the Cluster which then routes the traffic accordingly. That way I only use 3 Targets of my load balancer and my Cluster also controls the certs, so I don’t need to pay extra for more certs and load balancer recourses.
How could I accomplish that, and is this even possible when wanting a production grade/ready cluster?
You can use the Nginx ingress controller. Set it up using a single 'target,' possibly two for HTTP/S, and then internally route to your services using the Nginx controller.
@@AntonPutra Thanks! So my way wasn’t completely wrong. But as I understand this would now be a single point of failure, so how can I have HA Ingress Controller? Then the external load balancer can point to the 3 or more ingress controller.
Good to learn from you ❤
When we use the ingress, what should be the backend service type ? is it OK if its ClusterIP ?
Use ingress when you want to share one load balancer among multiple services, which is limited to HTTP/S only. Yes, most of the time you would use ClusterIP, except in some edge cases when you use the AWS Load Balancer Controller in instance mode.
@anton great work
Thank you Mengha!
Would you please share a tutorial for onpremise scenario, How can I deploy Ingress my onpremise kubernetes cluster?
Please please please 😊
For on premise use metallb to create load balancers, then use it to deploy ingress.
github.com/metallb/metallb
Thanks
what lesson number is this? can't fin the docs in github
I think I didn't commit any code for this lesson
Awesome!
Thank you!
Awesome :)
thanks!
GREAT
thank you!
Thanks Teacher
Subscribed and liked
thank you!
почему так сложно обяснить получше про LoadBalancer service? Ты говоришь, что он открывает под во вне, но совершенно непонятно, нужно навешивается реально какой-то балансировщик нагрузки или нет? например на 5:00 по смыслу да. а в остальном видео по смыслу нет.
I didn't understand the question. In the cloud, the LoadBalancer service creates a cloud load balancer and routes traffic to your app; on-premise, it depends on your configuration. Also, LoadBalancer can use node ports underneath, and if Kubernetes supports native networks (without virtual network plugins such as Flannel, Calico), it can route directly to your pod's IP address. (I don't have a Russian keyboard.)
@@AntonPutra Hey Anton. Thank you so much for answering! Now this seems to be clear for me.
What about using NodePort to aggregate different services under 1 load balancer using backends, instance-groups, url-maps, and HTTPS proxy?
Well that's how cloud providers used to do it behind the load balancer type service. Nowadays, they utilize IP mode and route traffic directly to the pod. On-premise, perhaps...
Anybody has an idea on the tools used to make this video (I mean animations and design) thank you !
I use adobe suite
@@AntonPutra Thank you for your answer, that looks very cool but also complicated I guess it's using after Effects right?
@@peace2941 Yes, I create graphics in Illustrator and animate them in After Effects.
@@AntonPutra Looks like a lot of effort, thank you, that looks really good !
@@peace2941 Thanks
➜ ~ kubectl -n kubernetes-dashboard port-forward svc/kubernetes-dashboard 8443:443
Error from server (NotFound): namespaces "kubernetes-dashboard" not found
Have you deployed the dashboard? Run "kubectl get ns" to list all namespaces.