As you may or may not be aware, all of our code is hosted on GitLab, and most of it is automatically tested using GitLab runners.
Currently, we do not use GitLab's continuous deployment features, which would enable code to be pushed directly from the repositories into production.
We are looking to change that.
here are some snippets of a conversation that I had today concerning kubernetes, which is our first set of conclusions on this initiative.
So, the conclusion is this:
We are going to attempt using Scaleway's Kubernetes service to enable CD for everything except full nodes.
We are still looking at some other options as well:
If you have thoughts on hosted Kubernetes systems, you should probably start running a witness! Not only should you start running a witness, you should advise us.
Links:
https://mlohr.com/kubernetes-cluster-on-hetzner-bare-metal-servers/
https://cloud.google.com/kubernetes-engine
https://www.ovhcloud.com/asia/public-cloud/kubernetes/
https://www.scaleway.com/en/kubernetes-kapsule/
https://www.scaleway.com/en/docs/how-to-use-the-gitlab-kubernetes-integration/
GOAL
ENABLE CONTINUOUS DEPLOYMENT
moving to continuous deployment will enable blurt to iterate more rapidly. Additionally, our final solution will be open sourced and documented so that community members wishing to run their own kubernetes systems will be able to do so with ease.