My friend, COBOL from 1969 still runs too. Jajaja. Is not argument, is archaeology!
What do you use for configuration management?
For anyone in EU, OVHcloud Gravelines to Marseille is 15ms. I use ansible-pull from GitHub for my personal nodes, no control machine. Works fine if you don't need push-now.
I administrate 15 servers for my association, all at OVHcloud Strasbourg and Gravelines. We use ansible since 2020. The YAML is not the problem, is the documentation in english only, my colleagues struggle. We translate internally.
I tried SaltStack for 20 nodes at AWS Dubai region. Very fast, but the documentation is worse than ansible. Also zero presence in middle-east forums, hard to find help. Back to ansible.
This. The YAML is a symptom. The real disease is that every module has 40 examples and zero explanation of *when* to use it.
Chicago here. I rack for colo clients who bring their own gear. Half use ansible, half use nothing. The ones using nothing have more downtime, but they blame the network first. Every time.
Manchester. I moved from puppet to ansible last year. Puppet was fine but the agent kept dying on Ubuntu 22.04, some systemd nonsense. Ansible has no agent to die. That alone won me.
COBOL doesn't ssh into 47 hosts with different kernel versions and cloud-init states. My bash handles that. Your ansible handles it slower and with more RAM.
99.97% is not 99.999%. Just saying.
Control node use 2GB RAM for 200 nodes. Is nothing. Your bash use your brain RAM, is expensive and not scalable. Jajaja.