burrowx - kafka offset lag monitor,stored by influxdb

A simple, lightweight kafka offset monitor, currently metrics stored by influxdb. Motivated by Burrow, but much faster and cleaner and more stable. burrowx is good integration with influxdb and grafana.

DemoView

Install

• By release binary

  Just click the release links

• By source build

$ go get github.com/sundy-li/burrowx
$ cd $GOPATH/src/github.com/sundy-li/burrowx
$ go build && go install

Running burrowx

## new workspace for burrowx
mkdir -p /data/app/burrowx 
## cd to the workspace
cd /data/app/burrowx 
## cp the files to here
cp $GOPATH/bin/burrowx ./
cp -rf $GOPATH/src/github.com/sundy-li/burrowx/config ./

## you should create the burrowx database in influxdb manually
## then modify server.json file config and run it 
./burrowx

Docker

A Docker file is available which builds this project on top of an Alpine Linux image.

1. Create your desired configuration files (server.yaml, logging.xml)
2. Run docker build -t burrowx . It will include the server.yaml and logging.xml and automatically start service.
3. Run the container on your favourite container platform

Test the data

• Create a new test topic

      bin/kafka-topics.sh --create  --zookeeper localhost:2181 --replication-factor 1 --partitions 8 --topic test_burrowx_topic
  

• Produce the data to the topic

  	for i in `seq 1 10000`;do echo "33" |  bin/kafka-console-producer.sh   --topic test_burrowx_topic --broker-list localhost:9092 ; sleep 1; done
  

• Create a consumer to consume the data

  $ pip install kafka

   from kafka import KafkaConsumer
   consumer = KafkaConsumer('test_burrowx_topic', group_id='my_group2')
   for msg in consumer:
   	print(msg)
  
   print("end")
  

Then you will find the data in the influxdb database burrowx .

Schema in influxdb

• cluster : cluster name
• topic : topic name
• consumer_group : group name
• partition : partition id
• logsize : partition logsize
• offsize : partition consumer offsize
• lag : partition consumer log

Query Example

SELECT sum("logsize") FROM "consumer_metrics" WHERE ("cluster" = 'your_cluster' AND "topic" = 'your_topic' AND "consumer_group" = 'your_consumer') AND $timeFilter  GROUP BY time(10s) 

SELECT sum("offsize") FROM "consumer_metrics" WHERE ("cluster" = 'your_cluster' AND "topic" = 'your_topic' AND "consumer_group" = 'your_consumer') AND $timeFilter  GROUP BY time(10s) 

SELECT sum("lag") FROM "consumer_metrics" WHERE ("cluster" = 'your_cluster' AND "topic" = 'your_topic' AND "consumer_group" = 'your_consumer') AND $timeFilter  GROUP BY time(10s) 

Features

• Light weight and extremely simple to use, metrics are stored in influxdb, and could be easily viewed on grafana
• Only support kafka version >= 0.9.X, which stores the consumer offsets in the topic __consumer_offsets,if you are using kafka 0.8.X, try my previous repo https://github.com/shunfei/Dcmonitor