Telemetry Showdown: Fluent Bit vs. OpenTelemetry Collector - A Comprehensive Benchmark Analysis
summary
Benchmark compares Fluent Bit 3 against OpenTelemetry Collector using real Hipster Shop + otel-demo load and Kepler high-cardinality exports; Fluent Bit handles logs far cheaper (Fluent Bit: ~60m CPU, ~15–20Mi RAM for logs) while Collector uses more CPU and memory for logs but provides full metrics/traces features. Collector shows a memory blowup when scraping Prometheus cumulative metrics with high cardinality (reaches ~10Gi and crashes) — root cause traced to Prometheus receiver conversion; adding the operator target-allocator offloads scraping and reduces Collector footprint to hundreds of Mi and ~100–200m CPU. Traces: Collector wins; metrics at scale: prefer Collector; logs-only: prefer Fluent Bit.
AI-written summary · about
chapters (7)
- 0:00Introduction
- 7:54Design comparison
- 18:42Logging plugins
- 23:40Metrics plugins
- 30:15Traces plugins
- 33:15Performance results
- 49:33Recommendations & Q&A
related talks
25mTelemetry Showdown: Fluent Bit Vs. OpenTelemetry Collector - a Comprehensive...
24mTelemetry Showdown: Fluent Bit Vs. OpenTelemetry Collector - A Comprehensive Benchma...
19mTelemetry Showdown: Fluent Bit vs. OpenTelemetry Collector - A Comprehensive Benchma...
07mFluent Bit vs. OpenTelemetry Collector - Why Not Both?
15mJourney from Fluent Bit, Fluentd and Prometheus to Open...
27mLog Support in OpenTelemetry
23mControlling Data Overhead with the OpenTelemetry Collector
06mFluent Bit V5