This fork is a modified version that support burrow 1.0.0 (Commit b8908854194b3cb59d34bb1805826855c1eb8fe6) There are a lot of changes on the endpoint version 2 into version 3
This datadog agent plugin will connect to instances of Burrow (https://github.com/linkedin/Burrow) and automatically get the offsets for every consumer and topic and push them into Datadog.
This plugin will push the offsets for all topics (except the offsets_topic) and consumers
for every kafka cluster it finds into Datadog as a metric.
The reported lags for each consumer group are also pushed up, along with the burrow consumer group status.
It will also push a check burrow.can_connect into Datadog to monitor the health of Burrow
Topic Offsets
kafka.topic.offsets- Per partition offsets of each topickafka.topic.offsets.total- Total offset of each topic
Consumer Offsets
kafka.consumer.offsets- Per partition offsets of each consumer groupskafka.consumer.offsets.total- Total offset of each consumer group
Lag
kafka.consumer.maxlag- Maximum lag of each consumer groupkafka.consumer.totallag- Total lag of each consumer groupkafka.consumer.partition_lag- Lag of each partition of a consumer group
The lag statuses are defined in burrow here: https://github.com/linkedin/Burrow/wiki/Consumer-Lag-Evaluation-Rules
Consumer group lag status:
kafka.consumer.lag_status.unknownkafka.consumer.lag_status.okkafka.consumer.lag_status.warnkafka.consumer.lag_status.errkafka.consumer.lag_status.stopkafka.consumer.lag_status.stallkafka.consumer.lag_status.rewind
Per Partition lag status of a consumer group:
kafka.consumer.partition_lag_status.unknownkafka.consumer.partition_lag_status.okkafka.consumer.partition_lag_status.warnkafka.consumer.partition_lag_status.errkafka.consumer.partition_lag_status.stopkafka.consumer.partition_lag_status.stallkafka.consumer.partition_lag_status.rewind
All metrics are tagged with cluster, topic, consumer and partition (where applicable) to make it easy to filter out in Datadog.
-
Start zookeeper and kafka up before Burrow
docker-compose up -d kafka zookeeper -
Start up burrow
docker-compose up burrow -
Populate some data into the kafka topic. Replace the IP address with whatever IP your kafka docker container is running as.
cat ./sample_messages | kafka-console-producer --broker-list 192.168.99.100:9092 --topic test-topic -
Consume some messages from the topic to get consumer offsets
docker-compose run datadog-agent /opt/datadog-agent/embedded/bin/python2.7 /test.py -
Execute the burrow check
docker-compose run datadog-agent /opt/datadog-agent/agent/agent.py check burrow