Hey Joonas,
I’ve finally had some time to try out Control Center, and I’d like to share my feedback with you.
After having difficulties with minikube and Podman in the beginning, I finally switched to Docker Desktop and have had a much better experiences with it.
The installation and configuration of Control Center went without any problems and, in my opinion, is very well done and really simple.
The navigation in Control Center was not immediately clear to me. Switching between the general settings and the settings for an app is not yet intuitive for me. Here are 2 examples:
- After I deployed an application, I wanted to customize the ReplicaSet and was looking for the right view. However, there doesn’t seem to be a suitable view for this in the context of the application. Only after I switched back to the general settings view for all applications, I was able to adjust the settings for an application. It is similar with the URL of the application. Unfortunately, the URL or link to the application cannot currently be found in the dashboard of the specific application, but only via the general settings view.
- Unfortunately, due to the current menu structure, it has also happened to me that I wanted to create roles, groups and users for a specific application, but accidentally did this for the entire Control Center because I was in the wrong menu context.
Another area in which I could imagine improvements is the display of errors for a deployment and the display of logs. If an error occurs during a deployment, I currently only see the status unavailable in the Control Center. I would like to see more details here, for example the specific condition or status of the pod or the events for a deployment or pod. Without this information, it is difficult to find out why a deployment might have failed. In such cases, I also installed the Kubernetes dashboard to give me a better overview of deployment errors.
The presentation of the logs for a pod could also be improved a little. On the one hand, the log statements in the grid component are not so easy to read due to the line spacing. On the other hand, it seems the logs statements are not persisted during a pod restart. Especially if a pod restarts frequently due to an error, it is difficult to view the logs then.
For the deployment of an application, I would like a few more setting options:
- It would be nice if you could configure the readiness, liveness and startup probe to be able to adjust the Kubernetes default values for certain applications, for example if an application simply needs significantly more time to start.
- Configuring the
ImagePullPolicywould also be helpful. Especially if you are working with local Docker images. See also Using local images with minikube - Octopus Deploy in this context. - For scaling the application, it would be nice if you could configure a Horizontal Pod Autoscaler as an alternative to the static replica set, for example. The deployment should be able to shut down pods that are no longer needed to save energy and money.
I am aware that all of these settings can also be made manually and without the help of Control Center. However, I wonder whether it is really worth considering for the target group of Control Center to make manual adjustments in the YAML file of deployments, or whether they would prefer to have a suitable interface in Control Center for this.
Last but not least, this error occurred from time to time when working with Control Center:
I had a lot of fun trying out Control Center, and I think that Control Center is on a very good path to simplify the management of Vaadin applications in Kubernetes.
I’m sure that the things I described here are familiar to you and are already in the backlog for further development of Control Center. Nevertheless, I wanted to take the time and write down my feedback because I know from my own experience as a product manager how difficult it can be to get feedback and how valuable feedback can be. So I hope this feedback is a bit useful for you.