Observability 2.0 - Much more than just logs,metrics and traces
summary
Claims that OpenTelemetry collector (receiver → processor → exporter) cuts downtime by standardizing telemetry and routing to multiple back ends, that eBPF surfaces kernel‑level traces for root cause, and that ML/LLMs detect anomalies, predict severity and summarize incidents. Describes a working ingest control pipeline that applies dynamic sampling filtering deduplication aggregation and edge processing to drop noise and save storage (e.g., avoid shipping logs to S3). Asserts faster recovery and click‑to‑scale automation.
AI-written summary · about
chapters (11)
- 0:00Introduction and Speaker Background
- 0:36Agenda Overview
- 0:48Evolution of Observability
- 3:57The Rise of Open Telemetry
- 5:25Understanding the Open Telemetry Collector
- 6:04Machine Learning in Observability
- 6:53Key Data Types in Observability 2.0
- 7:25Comparing Traditional and Modern Observability
- 7:58Cost Management and Smarter Data Pipelines
- 9:15AI-Driven Future of Observability
- 10:28Conclusion and Further Resources
related talks
22mUnified Observability
01mObservability
36mObservability For You and Me with OpenTelemetry
37mObservability For You And Me With Opentelemetry
50mGuide to Observability with OpenTelemetry
20mObservability and Open Telemetry
35mBeyond Logs, Metrics and Traces. The Data Analytics Paradigm Shift in Observability — Dotan Horovits
18mOpenTelemetry and Observability: What, Why, and Why Now?