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
What's still overhyped:
- Fully autonomous remediation without human approval (too risky in prod)
- AI writing your Helm charts from scratch (output needs heavy review)
Wrote a longer breakdown on my blog if anyone wants the full list with tool comparisons: https://infradecode.com
Curious what tools others are actually running in production — anything I missed?****
Top comments (0)