Question:
I assumed when I read the pitch, that I could spin up a managed K8s somewhere, like in Digital Ocean, and use this somehow. But after reading docs and comments, it sounds like this needs to manage my K8s for me?
I guess my question is:
1) When I spin up a "Cluster" on Hetzner, is that just dividing up a single machine, or is it a true K8s cluster that spans across multiple machines?
2) If I run this install script on another server, does it join the cluster, giving me true distributed servers to host the pods?
3) Is there a way to take an existing managed K8s and have Canine deploy to it?
Yeah so at the moment it kind of supports two options:
1. A single Hetzner VPS
2. An existing Kubernetes cluster.
I usually use #1 for staging / development apps, and then #2 for production apps. For #2, I manage the number of nodes on the Digital Ocean side, and kubernetes just magically reschedules my workload accordingly (also can turn on auto scaling).
I think the thing that you're getting at that is not supported is having Canine create a multi-node cluster directly within Hetzner.
There is a terraform to create a Kubernetes cluster from hetzner, but this isn't currently installed on Canine.
I'm not closed to trying it out, there were a few UI improvements I wanted to take a shot at first, but at the moment Canine assume's you have a cluster ready to go, or can help you walk through a K3s installation to a single VPS.
Oh! This is good news!
I was not asking about K8s on Hetzner per se. I was asking if I could spin up a managed cluster (on Digital Ocean, etc) and use this on it. It sounds like I can, which is great! I think I missed that in the docs.
Also, your docs on how K8s works look really good, and might be the most approachable docs I've seen on the subject. https://canine.gitbook.io/canine.sh/technical-details/kubern...
Question: I assumed when I read the pitch, that I could spin up a managed K8s somewhere, like in Digital Ocean, and use this somehow. But after reading docs and comments, it sounds like this needs to manage my K8s for me? I guess my question is: 1) When I spin up a "Cluster" on Hetzner, is that just dividing up a single machine, or is it a true K8s cluster that spans across multiple machines? 2) If I run this install script on another server, does it join the cluster, giving me true distributed servers to host the pods? 3) Is there a way to take an existing managed K8s and have Canine deploy to it?