Your Data is Made Powerful By Context (so stop destroying it already) (xpost)
This title could be clearer and more informative.Try out Clickbait Shieldfor free (5 uses left this month).
The three pillars model of observability (metrics, logs, traces stored in separate silos) destroys the relational context that makes telemetry data valuable. Data power grows combinatorially with each additional attribute, but siloing signals severs those relationships irreparably. AI-SRE agents are already bypassing traditional observability data in favor of richer, pre-digested telemetry with connective tissue intact. As agentic development accelerates change rates dramatically, precise production validation becomes critical — and that requires wide, context-rich, high-cardinality data stored together, not fragmented across pillars. AI-driven joins across silos cannot restore what was destroyed at write time.
Table of contents
Your observability data is self-destructing at write timeData is made valuable by relationshipsAI-SRE agents don’t seem to like three pillars dataYou can’t put Humpty back together againProduction is a very noisy placeAgents need speed, flexibility, context, and precision to validate in prodSort: