Hadoop HDFS DataNode¶

Collect HDFS datanode metrics.

Installation and Deployment¶

Since DataNode is developed in Java, metrics can be collected using the jmx-exporter method.

1. DataNode Configuration¶

1.1 Download jmx-exporter¶

Download URL: https://github.com/prometheus/jmx_exporter

1.2 Download jmx script¶

Download URL: https://github.com/lrwh/jmx-exporter/blob/main/hadoop-hdfs-datanode.yml

1.3 Adjust DataNode Startup Parameters¶

Add the following to the datanode startup parameters:

{JAVA_GC_ARGS} -javaagent:/opt/jmx/jmx_exporter-1.0.1.jar=localhost:17106:/opt/jmx/hadoop-hdfs-datanode.yml

1.4 Restart DataNode¶

2. DataKit Collector Configuration¶

2.1 Install DataKit¶

2.2 Configure the Collector¶

Since jmx-exporter can directly expose the metrics URL, it can be collected using the prom collector.

Navigate to the conf.d/prom directory under the DataKit installation directory, and copy prom.conf.sample to datanode.conf.

cp prom.conf.sample datanode.conf

Adjust the content of datanode.conf as follows:

  urls = ["http://localhost:17106/metrics"]
  source ="hdfs-datanode"
  [inputs.prom.tags]
    component = "hdfs-datanode" 
  interval = "10s"

Adjust other configurations as needed

, adjustment parameter descriptions:

• urls: jmx-exporter metrics URL, fill in the metrics URL exposed by the corresponding component here
• source: Collector alias, it is recommended to differentiate
• keep_exist_metric_name: Keep the metric name
• interval: Collection interval
• inputs.prom.tags: Add additional tags

3. Restart DataKit¶

Restart DataKit

Metrics¶

Hadoop Metrics¶

DataNode metrics are located under the Hadoop metrics, here we mainly introduce DataNode related metrics.