Create a Parser for Your New Data Source by Using the CLI

As an alternative to using the CCP Management module to parse your new data source, you can use the CLI.

  1. Determine the format of the new data source’s log entries, so you can parse them:
    1. Use ssh to access the host for the new data source.
    2. Look at the different log files and determine which to parse:
      sudo su - 
      cd /var/log/$NEW_DATASOURCE 
      The file you want is typically the access.log, but your data source might use a different name.
    3. Generate entries for the log that needs to be parsed so that you can see the format of the entries:
      timestamp | time elapsed | remotehost | code/status | bytes | method | URL rfc931 peerstatus/peerhost | type
  2. Create a Kafka topic for the new data source:
    1. Log in to $KAFKA_HOST as root.
    2. Create a Kafka topic with the same name as the new data source:
      --zookeeper $ZOOKEEPER_HOST:2181 --create --topic $NEW_DATASOURCE 
      --partitions 1 --replication-factor 1
    3. Verify your new topic by listing the Kafka topics:
      /usr/hdp/current/kafka-broker/bin/ --zookeeper $ZOOKEEPER_HOST:2181 --list
  3. Create a Grok statement file that defines the Grok expression for the log type you identified in Step 1.
    Refer to the Grok documentation for additional details.
  4. Save the Grok pattern and load it into Hadoop Distributed File System (HDFS) in a named location:
    1. Create a local file for the new data source:
      touch /tmp/$DATASOURCE
    2. Open $DATASOURCE and add the Grok pattern defined in Step 3:
      vi /tmp/$DATASOURCE
    3. Put the $DATASOURCE file into the HDFS directory where Metron stores its Grok parsers.
      Existing Grok parsers that ship with CCP are staged under /apps/metron/patterns:
      su - hdfs 
      hadoop fs -rmr /apparserps/metron/patterns/$DATASOURCE 
      hdfs dfs -put /tmp/$DATASOURCE /apps/metron/patterns/
  5. Define a parser configuration for the Metron Parsing Topology.
    1. As root, log into the host with CCP installed:
      ssh $CCP_HOST
    2. Create a $DATASOURCE parser configuration file at $METRON_HOME/config/zookeeper/parsers/$DATASOURCE.json:
      "parserClassName": "org.apache.metron.parsers.GrokParser",
      "filterClassName:": null,
      "writerClassName:": org.apache.metron,writer.kafka.KafkaWriter
      "sensorTopic": "$DATASOURCE",
      "readMetadata" : true,
      "mergeMetadata" : true,
      "rawMessageStrategy" : default
      "rawMessageStrategyConfig" : default
      "parserConfig": {   
         "grokPath": "/apps/metron/patterns/$DATASOURCE", 
         "patternLabel": "$DATASOURCE_DELIMITED", 
         "timestampField": "timestamp" 
      "fieldTransformations" : [ 
           "transformation" : "STELLAR" 
           ,"output" : [ "full_hostname", "domain_without_subdomains" ] 
           ,"config" : { 
                        "full_hostname" : "URL_TO_HOST(url)" 
                        ,"domain_without_subdomains" : 

      The name of the parser's class in the .jar file.

      The filter to use.
      This can be the fully qualified name of a class that implements the org.apache.metron.parsers.interfaces.MessageFilter<JSONObject> interface. Message filters enable you to ignore a set of messages by using custom logic. The existing implementation is STELLAR. The Stellar implementation enables you to apply a Stellar statement that returns a Boolean, which passes every message for which the statement returns true . The stellar statement is specified by the filter.query property in the parserConfig. For example, the following Stellar filter includes messages that contain a field1 field:
          "filterClassName" : "STELLAR"
         ,"parserConfig" : {
          "filter.query" : "exists(field1)"
      The class used to write messages after they have been parsed. Defaults to org.apache.metron.writer.kafka.KafkaWriter.

      The Kafka topic on which the telemetry is being streamed. If the topic is prefixed and suffixed by / then it is assumed to be a regex and will match any topic matching the pattern (for example, /bro.*/ matches bro_cust0, bro_cust1 and bro_cust2).


      A Boolean indicating whether to read metadata and make it available to field transformations (false by default).

      There are two types of metadata supported in CCP:

      • Environmental metadata about the whole system

        For example, if you have multiple Kafka topics being processed by one parser, you might want to tag the messages with the Kafka topic.

      • Custom metadata from an individual telemetry source that you might want to use within Metron

      A Boolean indicating whether to merge metadata with the message (false by default).

      If this property is set to true, then every metadata field becomes part of the messages and, consequently, is also available for field transformations.

      The strategy to use when reading the raw data and metadata.
      The following strategies are supported:
      Data is read directly from the kafka record value and metadata, if any, is read from the kafka record key. This strategy defaults to not reading metadata and not merging metadata. This is the default strategy.
      Data from kafka record value is presumed to be a JSON blob. One of these fields must contain the raw data to pass to the parser. All other fields should be considered metadata. The field containing the raw data is specified in the rawMessageStrategyConfig. Data held in the kafka key as well as the non-data fields in the JSON blob passed into the kafka value are considered metadata. Note that the exception to this is that any original_string field is inherited from the envelope data so that the original string contains the envelope data. If you do not prefer this behavior, remove this field from the envelope data.
      The raw message strategy configuration map.
      The following configurations are strategy dependent:
      metadataPrefix defines the key prefix for metadata (default is metron.metadata).
      metadataPrefix defines the key prefix for metadata (default is metron.metadata)
      messageField defines the field from the envelope to use as the data. All other fields are considered metadata.

      A JSON map defining the parser implementation configuration.

      This parameter also includes batch sizing and timeout settings for writer configuration. If you do not define these properties, the system uses their default values.

      • grokPath - The path in HDFS (or in the Jar) to the grok statement. By default attempts to load from HDFS, then falls back to the classpath, and finally throws an exception if unable to load a pattern.
      • batchSize - Number of records to batch together before sending to the writer. Default is 15.
      • patternLabel - The name of the Grok statement that defines the pattern of the Grok expression.
      • timestampField - Use the timestamp value to ensure that the system uses the event time rather than the system time.
          "parserConfig": {   
         "grokPath": "/apps/metron/patterns/$DATASOURCE", 
         "patternLabel": "$DATASOURCE_DELIMITED", 
         "timestampField": "timestamp"
      In addition, you can override settings for the kafka writer within the parserConfig file.

      An array of complex objects representing the transformations to be performed on the message generated from the parser before writing to the Kafka topic.

      In this example, the Grok parser is designed to extract the URL, but the only information that you need is the domain (or even the domain without subdomains). To obtain this, you can use the Stellar Field Transformation (under the fieldTransformations element). The Stellar Field Transformation enables you to use the Stellar DSL (Domain Specific Language) to define extra transformations to be performed on the messages flowing through the topology.


      The security protocol to use for reading from kafka (this is a string). This can be overridden on the command line and also specified in the spout config via the security.protocol key. If both are specified, then they are merged and the CLI will take precedence. If multiple sensors are used, any non "PLAINTEXT" value will be used.


      Cache config for stellar field transformations. This configures a least frequently used cache. This is a map with the following keys. If not explicitly configured (the default), then no cache will be used.

        • stellar.cache.maxSize - The maximum number of elements in the cache. Default is to not use a cache.
        • stellar.cache.maxTimeRetain - The maximum amount of time an element is kept in the cache (in minutes). Default is to not use a cache.
      Example of a cache config to contain at max 20000 stellar expressions for at most 20 minutes:
        "cacheConfig" : {
          "stellar.cache.maxSize" : 20000,
          "stellar.cache.maxTimeRetain" : 20
    3. If you want to set the grok parser to use the current year in its timestamp, add the following information to the transformations function in the datasource json file:
      "fieldTransformations" : [
                "transformation" : "STELLAR"
                ,"output" : [ "timestamp"]
                ,"config" : {
                          "timestamp”: “TO_EPOCH_TIMESTAMP(FORMAT(‘%s %d’, timestamp_str , YEAR()), ‘MMM dd HH:mm:ss:yyyy’)”
      For example, the datasource json file would change to:
      "fieldTransformations" : [
                "transformation" : "STELLAR"
                ,"output" : [ "full_hostname", "domain_without_subdomains" , "timestamp"]
                ,"config" : {
                          "full_hostname" : "URL_TO_HOST(url)"
                          ,"domain_without_subdomains" :
                          ,”timestamp”: “TO_EPOCH_TIMESTAMP(FORMAT(‘%s %d’, timestamp_str , YEAR()), ‘MMM dd HH:mm:ss:yyyy’)”
    4. Use the following script to upload configurations to Apache ZooKeeper:
      $METRON_HOME/bin/ --mode PUSH -i $METRON_HOME/config/zookeeper -z $ZOOKEEPER_HOST:2181
  6. Deploy the new parser topology to the cluster:
    If you want to deploy multiple parsers on one topology, refer to Creating Multiple Parsers on One Topology.
    1. Log in to the host that has Metron installed as root user.
    2. Deploy the new parser topology:
    3. Use the Apache Storm UI to verify that the new topology is listed and that it has no errors.
    This new data source processor topology ingests from the $DATASOURCE Kafka topic that you created earlier and then parses the event with the CCP Grok framework using the Grok pattern defined earlier.