I'm designing a portfolio deploy-and-destroy project, and I'm half wonder if I hate myself enough to make it multicloud
I don't even *know* what I want to deploy at the end of this. Apparently I just like making my life difficult.
Multicloud isn't inherently hard, I'll point out, it's the cost-related issues of deploying a multicloud setup. I'd probably be delegated to AWS, GCP, and *maaaaybe* Linode (I don't think Oracle or Azure have a free tier but I haven't checked in a hot minute)
I know I'll be using:
- Ansible for provisioning at-scale
- Prometheus for systems monitoring
- Elastic for logs mgmt
- Github actions for my CI (I need to shell out for Gitlab and it's just not possible right now)
All of it's application specific but I'm trying to work on some of my weaker areas
@KAutumnrain good stack 👌 any orchestration and containers system to dev, QA things out ?
I'm trying to think of whether or not I should use k8s or a docker swarm, for sure. I think either/or would be a good reinforcement and learning experience.
@KAutumnrain k8s is my choice usually (cause I know it lol). My main devops challenge in all if this usually is how to convert old school projects and their tech teams to it (they have their habits and its a learning curve for them.). Maybe the key is to let them be until they are ready to go into full refactoring mode !
@KAutumnrain oh yeah I understand. Here the business decisions makes us the same (not using the whole devsecops concepts hehe). Too bad but eventually you will get this one client who will want it all ! 👍
@Guillaume
HAH maybe.
k8s is typically my thing as well. but as of current at work we haven't exactly been getting contracts of really anything-at-scale. I find myself using a lot of my skills in parts than as one unified experience. Trying to keep things sharp and hopefully move on to something a little bit more sustainable for myself.