Let’s say engineers want to investigate a production issue. Today, they typically face a maze of dashboards they’ve never seen, tools they haven’t been trained on, and terms like ‘PromQL’, ‘context propagation’, and ‘cardinality’- which might as well be a foreign language.

We’ve built an observability industry so complex that most engineers can’t observe their own code. The tools exist, the data exists— but the accessibility doesn’t.

How do we make production insights understandable to the people who build the product, not just the specialists who run it? When your engineers can’t troubleshoot independently, every incident becomes a bottleneck— and that’s an organizational problem, not just a tooling one.

Let’s discuss with Jon Reeve, co-founder and CPO of ControlTheory, an Austin-based AI-first observability startup developing practical AI for SREs. I met Jon at KubeCon Atlanta, and I think his philosophy and approach can help cut through the ‘complexity cult’ that dominates observability today.

Links:

• ControlTheory
• Gonzo
• Slack Community