Enabling External JMX Tools

Data Collector uses JMX metrics to generate the graphical display of the status of a running pipeline. You can provide the same JMX metrics to external tools if desired.

Information provided by JMX metrics includes pipeline details like a histogram for the number of error records per batch or the amount of memory the pipeline uses. Stage-related details are also provided, such as the number of output records or stage errors. Some stages have stage-related custom metrics.

The following Java environment variables expose the Data Collector JMX metrics on a specified port, allowing integration with external tools:
  • com.sun.management.jmxremote
  • com.sun.management.jmxremote.port=<port_number>
  • com.sun.management.jmxremote.local.only=<true | false>
  • com.sun.management.jmxremote.authenticate=<true | false>
  • com.sun.management.jmxremote.ssl=<true | false>

You can pass the variables in the command line as part of the SDC_JAVA_OPTS environment variable. Or, you can add the variables to the SDC_JAVA_OPTS environment variable in the Data Collector environment file.

For example, the following set of variables passes JMX metrics through port 3333:

export SDC_JAVA_OPTS="-Dcom.sun.management.jmxremote \
-Dcom.sun.management.jmxremote.port=3333 \
-Dcom.sun.management.jmxremote.local.only=false \
-Dcom.sun.management.jmxremote.authenticate=false \
-Dcom.sun.management.jmxremote.ssl=false"

Viewing JMX Metrics in External Tools

You can view the Data Collector JMX metrics in external tools. The Data Collector JMX metric names all begin with "sdc.pipeline."

Data Collector JVM metrics use the following naming pattern:

sdc.pipeline.<pipeline name>.<pipeline revision>.<category: pipeline|stage|custom>.\
[<stage library>_<library revision>].<metric name>.<metric type>

For example, the following is a batch count meter for the first revision of a pipeline named Write to Kafka:

sdc.pipeline.Write to Kafka.0.pipeline.batchCount.meter
The following metric is a counter for the memory consumed by the File Tail origin in the same Write to Kafka pipeline:
sdc.pipeline.Write to Kafka.0.stage.\
com_streamsets_pipeline_stage_origin_logtail_FileTailDSource_1.memoryConsumed.counter

Custom Metrics

Data Collector provides custom metrics for some stages. When a pipeline includes the stages below, you can view custom metrics for the stages in the Data Collector console in Monitor mode or when you view JMX metrics using an external tool:

File Tail origin
In addition to the standard metrics available for origins, File Tail provides the following custom metrics:
  • Offset Lag - The amount of data remaining in the file being read. This metric displays in external tools as follows:
    sdc.pipeline.<pipeline name>.<pipeline revision>.custom.\
    com_streamsets_pipeline_stage_origin_logtail_FileTailDSource_\
    <library version>.offsets.lag.<file path>.counter
  • Pending Files - The number of files in the directory that still need to be read. This metric displays in external tools as follows:
    sdc.pipeline.<pipeline name>.<pipeline revision>.custom.\
    com_streamsets_pipeline_stage_origin_logtail_FileTailDSource_\
    <library version>.pending.files.<file path>.counter
Amazon S3 destination
In addition to the standard metrics available for origins, Amazon S3 provides the following custom metrics:
  • Transfer Rate KB Meter - Displays the transfer rate in KB. Appears when the destination writes whole files to the destination system with the whole file data format. The counter displays in external tools as follows:
    sdc.pipeline.<pipeline name>.<pipeline revision>.custom.\
    com_streamsets_pipeline_stage_destination_s3_\
    AmazonS3DTarget_<library version>.transferRateKb.meter
Hadoop FS destination
In addition to the standard metrics available for origins, Hadoop FS provides the following custom metrics:
  • Late Records meter and counter - The number of late records written to HDFS. The counter displays in external tools as follows:
    sdc.pipeline.<pipeline name>.<pipeline revision>.custom.\
    com_streamsets_pipeline_stage_destination_HdfsTarget_\ 
    HDFSDTarget_<library version>.lateRecords.<counter | metric>
  • To HDFS Records meter and counter. The number of records written to HDFS. The counter displays in external tools as follows:
    sdc.pipeline.<pipeline name>.<pipeline revision>.custom.\
    com_streamsets_pipeline_stage_destination_HdfsTarget_\
    HDFSDTarget_<library version>.hdfsRecords.<counter | metric> 
  • Transfer Rate KB Meter - Displays the transfer rate in KB. Appears when the destination writes whole files to the destination system with the whole file data format. The counter displays in external tools as follows:
    sdc.pipeline.<pipeline name>.<pipeline revision>.custom.\
    com_streamsets_pipeline_stage_destination_HdfsTarget_HDFSDTarget_\
    <library version>.transferRateKb.meter
Local FS destination
In addition to the standard metrics available for origins, Local FS provides the following custom metrics:
  • Late Records meter and counter - The number of late records written to the local file system. The counter displays in external tools as follows:
    sdc.pipeline.<pipeline name>.<pipeline revision>.custom.\
    com_streamsets_pipeline_stage_destination_localfilesystem_\
    LocalFileSystemDTarget_<library version>.lateRecords.\
    <counter | metric>
  • To HDFS Records meter and counter. The number of records written to the local file system. The counter displays in external tools as follows:
    sdc.pipeline.<pipeline name>.<pipeline revision>.custom.\
    com_streamsets_pipeline_stage_destination_localfilesystem_\
    LocalFileSystemDTarget_<library version>.hdfsRecords.\
    <counter | metric>
  • Transfer Rate KB Meter - Displays the transfer rate in KB. Appears when the destination writes whole files to the destination system with the whole file data format. The counter displays in external tools as follows:
    sdc.pipeline.<pipeline name>.<pipeline revision>.custom.\
    com_streamsets_pipeline_stage_destination_localfilesystem_\
    LocalFileSystemDTarget_<library version>.transferRateKb.meter
MapR FS destination
In addition to the standard metrics available for origins, MapR FS provides the following custom metrics:
  • Late Records meter and counter - The number of late records written to MapR FS. The counter displays in external tools as follows:
    sdc.pipeline.<pipeline name>.<pipeline revision>.custom.\
    com_streamsets_pipeline_stage_destination_marpfs_\
    MaprFSDTarget_<library version>.lateRecords.<counter | metric>
  • To HDFS Records meter and counter. The number of records written to MapR FS. The counter displays in external tools as follows:
    sdc.pipeline.<pipeline name>.<pipeline revision>.custom.\
    com_streamsets_pipeline_stage_destination_marpfs_\
    MaprFSDTarget_<library version>.hdfsRecords.<counter | metric>
  • Transfer Rate KB Meter - Displays the transfer rate in KB. Appears when the destination writes whole files to the destination system with the whole file data format. The counter displays in external tools as follows:
    sdc.pipeline.<pipeline name>.<pipeline revision>.custom.\
    com_streamsets_pipeline_stage_destination_marpfs_MaprFSDTarget_\
    <library version>.transferRateKb.meter