Anyone interested to build web and cli for orchestration and monitoring based on LXD

1

I am interested in building a simple web and cli interface for orchestration, monitoring and scaling of LXD based cluster. If interested let me know through this forum.

I have been looking at Kubernetes for a while and I feel its unnecessary complex. Technology with modern progress should make things easier but it seems with new acronyms everyday like Kubectl, Helm, ISTIO and on and on, its making things harder then it has to be. Containers were suppose to make life easier for managing application with dependencies, now it seems deployment itself is harder then the application. I think today with kubernetes it will take longer to build a CI/CD and deployment pipeline then write and manage application on baremetal cloud.

Personally I like LXD as its easier and the concepts are cleaner and by default more secure compared to docker baased containers, so thought of this.

2

Potentially interested in helping to test this…

3

Hey take a look at my web interface here, I have deployments that take collections cloud configs and deploy them

2 Likes
4

It’s cool, I need to get into clusters but this looks like a great way to go…

5

We are looking for a simple orchestration solution for LXD/LXC to enable auto scaling of apps/containers. Just started to get my feet wet in LXD/LXC. Been a Docker and Kubernetes engineer for the past 3-4 years and have a client that is looking for GUI, clustering and auto scaling in GCP. I will post a new entry regarding this…just reaching out on this post. Thx.

1 Like
6

In short: I haven’ seen an open source native LXD scaling option yet - Look at LXE

Scaling is hard in LXD (in my opinion and understanding) because they don’t offer CPU load averages over the API (most apps ive seen aren’t bound by memory) - I have a pull request Ive been neglecting for a while that introduces the capability to record these load averages (and more) in my program.

Once load avgs are combined with memory data trying to scale apps should be more practical, in my program (along time ago) I wrote deployments which acts away to group cloud-configs (effectively docker files) together and run containers (I.E give me X of XYZ type containers, phone home when ready ETC) in preparation of my “cracking” the CPU load avgs problem - the intent being to scale deployments - but it took me like a year (because im lazy) and deployments has probably fallen into a disgusting state since.

If your more familiar with the Kubernetes you may want to look at the LXE shim (I have no experience but its been around a while)