Leveraging OpenTelemetry for Your Prometheus Pipeline
summary
OpenTelemetry Metrics is GA and promises Prometheus compatibility; the Collector can ingest via the Prometheus receiver (wraps full Prom config), the simple Prometheus receiver (single-endpoint, no service discovery), or the deprecated exec receiver, and exports via a /metrics Prometheus exporter or, recommended for scale, the remote_write exporter (and a Google Managed Prometheus variant). Gotchas: dots become underscores, resource labels map into job/instance/target_info (requiring joins), and exponential histograms aren’t fully supported yet. Use processors to add labels or migrate instrumentation to OTLP SDKs.
AI-written summary · about
chapters (8)
- 0:00Introduction
- 0:45Prometheus
- 1:30OpenTelemetry
- 4:46Prometheus Receiver
- 8:43Prometheus Exporter
- 13:35Gotchas
- 16:53Use Cases
- 19:45Questions
related talks
22mBuilding Prometheus Metrics Support in OpenTelemetry
27mThe State of Prometheus and OpenTelemetry Interoperability
13mConnecting Prometheus and OpenTelemetry Data for Faster Troublesho...
36mShould I use OTel (collectors), or is Prometheus good enough?
28mOpenTelemetry developments in Prometheus over the last year
41mNavigating the Open Source Observability Landscape
28mPrometheus 3.0 Overview
30mOpenTelemetry 101