OpenTelemetry Metrics: observability’s holy grail
summary
OpenTelemetry Metrics presents a full Java API/SDK plus OTLP and a lightweight Collector that together replace Micrometer for many teams, provide views override attributes to drop high‑cardinality labels and change histogram buckets at startup, and expose data via a bundled Prometheus /metrics endpoint or periodic readers with exporters (otlp gRPC or HTTP). The SDK adds semantic conventions and temporal (cumulative vs delta) aggregations; a reported optimization cut metric memory allocations by ~98% while meeting <5 ms latency needs for Apache Pulsar.
AI-written summary · about
chapters (8)
- 0:00Introduction and speaker background
- 6:10Why OpenTelemetry metrics
- 9:21OpenTelemetry project overview
- 14:42OpenTelemetry metrics API basics
- 23:32OpenTelemetry metrics SDK deep dive
- 33:15Views and attribute management
- 37:47Summary and future potential
- 42:02Q&A session
related talks
42mOpenTelemetry Metrics 101
20mObservability and Open Telemetry
36mObservability For You and Me with OpenTelemetry
32mOpenTelemetry feat. logs, metrics and traces
18mOpenTelemetry and Observability: What, Why, and Why Now?
37mObservability For You And Me With Opentelemetry
30mOpenTelemetry 101
50mOpen Telemetry: The Rise of Open Standards in Observability