After a decade building large-scale systems at Google, Datadog, and Meta, I’ve noticed the same pattern repeat: observability keeps getting louder, costlier, and less useful.
We’re drowning in telemetry but starving for insight. The industry incentives are misaligned: they reward ingestion and storage, not intelligence.
I recently started an open, collective movement called omji.ai to explore a fundamental shift: measuring insight per dollar of telemetry stored. We need to push vendors and internal teams toward intelligence, not ingestion.
I’m curious to hear from folks facing the pain - how do we fix this? We need practical, non-obvious ideas.
1. What technical or economic levers would actually shift the industry's focus from volume to intelligence?
2. Has anyone in a large organization tried benchmarking observability systems based on insight (e.g., MTTR impact) vs. telemetry cost?
3. How could open collaboration (tools, standards, benchmarks) make this practical for every engineering team?
Full background post: https://www.linkedin.com/pulse/observability-broken-lets-cha...
