I've been testing a bunch of AI tools in my Kubernetes workflow over the past few months and wanted to share what's genuinely changed my day-to-day vs what's just marketing noise.
What's actually working:
K8sGPT — scans your cluster and explains issues in plain English. Saved me a lot of time on pod crash debugging. Open source, worth trying.
AI-assisted incident triage — tools that correlate logs + metrics and surface root cause faster than manual grep-ing through Kibana
Natural language infra provisioning — still early but some teams are running Terraform via prompts in CI pipelines
