Bill Hineline & Margaret Dawson, Chronosphere | Red Hat Summit 2025
summary
Chronosphere demonstrates an observability stack that controls ingestion at scale, uses utility scoring to drop roughly 80%+ of unused telemetry while archiving the rest, and reports four nines availability across massive workloads. It integrates OpenTelemetry and Prometheus-style signals, pushes SLO/error‑budget workflows, and surfaces correlated GPU and LLM metrics so teams can diagnose hallucinations or GPU contention. The company also pushes a commercial shift: pay for what you keep rather than ingest volume billing.
AI-written summary · about
chapters (10)
- 0:00Welcome and introduction
- 1:12Why modern apps are complex
- 2:35Why APM no longer works
- 5:00Modern observability strategy
- 8:34Changing personas in observability
- 10:42Dangers of too much data
- 13:15Alert fatigue and trust
- 16:00AI's impact on observability
- 18:55Future of observability pricing
- 20:00Closing remarks and thanks
