Writing

Building an autonomous SRE Agent taught us that more context produces better root-cause analysis. Rather than centralizing more telemetry, the answer is to move intelligence closer to where the data lives.

An overview of a three-part series on Kubernetes cost management: where the spend comes from, how to optimize it as a feedback loop, and how to run that loop as a FinOps practice with SRE Agents inside human-set boundaries.

Once you can find optimization opportunities, human attention becomes the bottleneck. This piece applies the Crawl, Walk, Run maturity model to Kubernetes and looks at where SRE Agents fit.

Kubernetes cost optimization fails when it is treated as a one-time cleanup. This piece covers attribution, baselines, rightsizing, cluster optimization and validation as one continuous loop.

AI-assisted coding didn't take the satisfaction out of building for me. It moved it, from writing code to reasoning about the system that code belongs to.

A breakdown of where Kubernetes spend actually comes from, with real examples of rightsizing and hidden network costs, and why most organizations can't answer basic questions about their bill before they try to cut it.