Improving Kafka monitoring with native Prometheu...
summary
Strimzi implements a native Prometheus metrics reporter that removes JMX exporter and exposes Prometheus-format metrics over HTTP (default port 9404), filtering metrics early to avoid collecting thousands first. Brokers and controllers are supported now; Connect and MirrorMaker 2 are in progress and the HTTP bridge gains config to choose reporter or exporter. Demo shows a configurable allow list and a single-metric scrape, performance tests show a flat scrape latency vs exporter, and the reporter supports dynamic reconfig without restart.
AI-written summary · about
chapters (11)
- 0:00Kafka monitoring with Prometheus
- 1:04Why monitoring matters
- 3:13Prometheus metrics & Grafana
- 5:42Current pipeline with JMX
- 8:29Native Prometheus reporter
- 12:47Live demo of reporter
- 17:10Integration status & plans
- 18:16HTTP bridge metrics
- 20:24Reporter performance
- 21:27Future features
- 25:00Q&A
related talks
33mObservability-First Kafka: Engineering Visibility at Scale
29m“If you Don’t Monitor your Infrastructure, you Don’t Own it!” Regain Control Thanks to Prometheus
01h 03mWhat’s new in the Prometheus ecosystem?
32mDemocratizing Application Monitoring with Prometheus
36mKafka Monitoring: What Matters!
29mEnable Serverless Metrics in Apache OpenWhisk on Kubernetes with Prometheus
29mMastering Kafka Workload Balancing With Strimzi’s Cruise Control Integration
34mMonitor My Socks: Using Prometheus in a Polyglot Open Source Microservices