Ideas
Half-baked thoughts, rabbit holes, and things I'm exploring. Not everything here will ship — that's the point.
exploring
Platform-as-Product Metrics
What if we treated internal platforms like SaaS products and tracked NPS, adoption funnels, and time-to-value for developer workflows?
building
Ephemeral Preview Environments
Spin up a full-stack preview environment per PR using Crossplane + ArgoCD. Destroy on merge. Zero manual steps.
thinking
Infrastructure Decision Records
Like ADRs but specifically for infrastructure decisions. Why did we pick this region? Why this instance type? Embed context in the codebase.
thinking
On-Call as a Service
What would it look like to make on-call rotation setup, escalation policies, and runbook linking fully self-service for product teams?
exploring
Cost-Aware Autoscaling
HPA decisions that factor in spot instance availability and cost curves, not just CPU/memory utilization.