The Hashicorp suite vs Docker/K8s will be a big one. My first impression was that containerizing everything was just the obviously superior approach, but then I had to actually deploy a job to K8s and I discovered that the technology has serious stability issues. I still like the idea, but when I’m being paid to write and deploy automation that just works I can’t afford to “follow my heart”, so to speak.
To be fair, I’m in a situation where I can control the entire stack if I go with Terraform, whereas the K8s clusters are controlled by a different team. Still, the K8s cluster node was the point of failure. (It suddenly stopped being able to resolve AWS service endpoints.)
To be fair, I’m in a situation where I can control the entire stack if I go with Terraform, whereas the K8s clusters are controlled by a different team. Still, the K8s cluster node was the point of failure. (It suddenly stopped being able to resolve AWS service endpoints.)