Kafka Lag & Brokers Lite (kafka_exporter)

Kafka consumer lag by group, topic and partition, under-replicated partitions, broker count, produce rate versus consume rate and consumer group members. Pick job, consumer group, topic and top N at the top. Requires Prometheus scraping kafka_exporter (danielqsj/kafka_exporter). Free alert rules and Telegram routing: https://devkit-shop.pages.dev/free/

Find the stuck consumer fast. The dashboard shows consumer lag per group, per topic and per partition, so one stalled partition does not hide behind a small total.

Cluster health is covered by the broker count, under-replicated partitions and missing in-sync replicas per topic. Throughput panels compare the produce rate (from partition end offsets) with the consume rate, and the members panel shows when a group has lost its consumers.

Variables: job, consumer group, topic and top N. The data source is chosen at import time.

Requirements

Grafana 10.0 or newer
Prometheus scraping kafka_exporter (danielqsj/kafka_exporter): kafka_brokers, kafka_consumergroup_lag, kafka_consumergroup_lag_sum, kafka_consumergroup_current_offset_sum, kafka_consumergroup_members, kafka_topic_partitions, kafka_topic_partition_current_offset, kafka_topic_partition_replicas, kafka_topic_partition_in_sync_replica, kafka_topic_partition_under_replicated_partition