What are the options for organizing the assembly, deployment and monitoring of applications in kubernetes?

Question

Software Deployment

Good afternoon,

I am looking for tools to organize the assembly of projects in docker images, put them in a certain registry, upload them to a kubernetes cluster and monitor kubernetes resources, preferably from a web ui, so that I don’t have to go to the server and work with the terminal.

Now we have our own server with Gitlab, it is configured to build images, work with registry and roll out to kubernetes.

But there is no monitoring.

Initially, I wanted to do everything inside Gitlab, but after laying it out in cubes, you have to wait until it rises and monitor whether the services have risen. So far, it has been implemented by separate manual tasks, but this is not convenient.

Therefore, there are 2 options
- find the ability to monitor the layout and resources in Gitlab
- limit work in Gitlab to only building images and placing them in the registry and find a separate tool for laying out and monitoring cubes

In general, this is the whole question - what are the options for implementing these two approaches? Ideally, I would like to implement the second option so that there is work with cubes in 1 place, and not in 2 (gitlab and monitor)

PS Another important point is to be able to monitor pod logs. I would very much like to. Or just run your commands against the cluster and see the output in the web UI.

Answer 1

Here you have made a mess in the question of CI, CD, monitoring and logging!
If you want in one place - Gitlab has integration with Kubernetes. Unfortunately, I'm not sure if it will cover your wishes 100%, but it seems that it could do basic monitoring and display of logs https://docs.gitlab.com/ee/user/clusters/agent/ind...
If you separate, then already you will have to build on the desired deployment solution. I will describe using my own example - the GitOps approach suits me best, and I chose to roll out using helm and ArgoCD - there is minimal monitoring and logging in one place.
Centralized monitoring of services is organized with the support of Prometheus / Prometheus Operator.
Logging in ELK with collection of logs via filebeat.
If you need a visual perception of the state of the cluster, I use Lens.

Answer 2

There is a convenient admin panel for the cuber Rancher, which allows you to tune with the mouse, and not with the console. There you can also add monitoring and logging. It is better to put it not in a cluster, but outside, because if the cluster is well, then the admin panel too.