Skip to content

CRAIG 5.1.2

Below is a summary of the JIRA issues addressed in the CRAIG-5.1.2 release of Punchplatform. For full documentation of the release, a guide to get started, and information about the project, see the Punchplatform project site.

Note about upgrades: Please carefully review the upgrade documentation for this release thoroughly before upgrading your cluster. The upgrade notes discuss any critical information about incompatibilities and breaking changes, performance changes, and any other changes that might impact your production deployment of Punchplatform.

The documentation for the most recent release can be found at

Release summary

This release brings major fixes to version 5.1.1. It also brings deployment improvements and provides MapD and Azure bolts.

Important changes

Please read Migration Guide from 5.1.1 to 5.1.2

Release notes

Main subjects

  • [PP-2955] - First Level Integration of DeepLearning in the Punch
  • [PP-2851] - Ergonomics and functionality for PML

New feature

  • [PP-2882] - Support Hjson for punchlets
  • [PP-2999] - Data science tools with ShowMultiClassMetrics node
  • [PP-2919] - Allow description field per PML Node
  • [PP-2918] - Allow to "open" a PML that has been saved to Elasticsearch
  • [PP-2909] - MLLib stage model file save, reload and reuse
  • [PP-2870] - Support custom node in PML UI


  • [PP-2854] - Upgrade Elastic components from 6.4.0 to 6.5.4
  • [PP-2858] - Upgrade Spark from 2.3.2 to 2.4.0
  • [PP-2939] - Siddhi version and API upgrade from 3.1.3 to 4.3.17
  • [PP-2866] - Azure Spout: avoid duplicate polling on azure nsg and add regex polling based on first letter of each file
  • [PP-2880] - ipmatch punch operator must support attachments
  • [PP-2901] - User-friendly errors/exceptions on PML configuration
  • [PP-2912] - Detect at PML launch time missing parameters
  • [PP-2416] - Improve Kibana Punch plugin usability
  • [PP-2417] - Standardization of errors for commands used by HMI
  • [PP-2920] - Improved content for PML Nodes description
  • [PP-2916] - Protect edited PML graph against unwanted closure
  • [PP-2700] - Minor migration doc from 5.0.0 to 5.1.0
  • [PP-2779] - Remove exception_catcher reference (legacy in Craig)


  • [PP-2247] - Some Storm metrics are likely the same
  • [PP-2276] - One node data management has an invalid metric reporter section
  • [PP-2634] - Wrong error message for Kafka Spout
  • [PP-2689] - Unflat tuple does not work when nestedSeparator settings is not a dot
  • [PP-2839] - Date operator default year make timestamps loop over the year of punchlet start.
  • [PP-2843] - log injector crashes when reading kafka topics using earliest strategy
  • [PP-2844] - file bolt cannot work without Elasticsearch
  • [PP-2845] - wrong error message when starting a channel with a wrong
  • [PP-2849] - kafka pml nodes node working nor documented
  • [PP-2861] - creates zookeeper.log file anywhere
  • [PP-2864] - TrainValidationSplit PML Stage cannot be configured using Kibana plugin
  • [PP-2874] - MLLib Validator stages cannot be used in PML
  • [PP-2879] - Fail to deploy spark
  • [PP-2881] - standard parser IBM_IPS unit_2 test fail
  • [PP-2902] - Month Jan/Jun/Jul are not recognize by syslogHeader()
  • [PP-2906] - activating spark logger does not work as expected
  • [PP-2921] - Repair PML Nodes documentation to HMI to ease documentation effort
  • [PP-2923] - Handle csv header column in file_csv_generator(input) node
  • [PP-2934] - pml guava jar conflicts
  • [PP-2942] - no and punchplatform-deployment.settings in 1-nodes
  • [PP-2944] - GeneratorSpout not working for some case
  • [PP-2953] - one of the mode does not work
  • [PP-2954] - punchplatform-analytics regression wrt elasticsearch
  • [PP-2958] - Mistyped tag in PML leads to unclear error
  • [PP-2965] - Invalid PML jar because of security exception
  • [PP-2974] - Kafka monitoring service fails
  • [PP-2982] - Putconf not working if PUNCHPLATFORM_CONF_DIR is not normalized
  • [PP-2984] - User-friendly error when missing or invalid tenant/etc/conf.json
  • [PP-2985] - Jar conflict between punch and spark commons.lang3
  • [PP-2986] - Standalone Elasticsearch not listening on 9200
  • [PP-2994] - OOM during the channel launching

Internal Task

  • [PP-2516] - ISettingsMap API robustness
  • [PP-2765] - Move online debugger to pp-gui
  • [PP-2800] - Create a contribute section in documentation
  • [PP-2860] - Unit test for Azure Spout blob storage
  • [PP-2862] - Put Zookeeper log in the right place
  • [PP-2867] - Improve pml files ergonomy
  • [PP-2905] - add --verbose option to avoid print or show nodes
  • [PP-2907] - speed up punch java apps startup time
  • [PP-2917] - Improve forms and graph for kibana plugin
  • [PP-2930] - --inline-json option must be possibly inline hjson
  • [PP-2957] - ensure the operator has a working environment without git
  • [PP-2975] - new csv and json output node for pml
  • [PP-2976] - 3 pml job configuration
  • [PP-2983] - Extracting data with PML and the FileOutput node
  • [PP-2989] - let the deployer populate zookeeper with initial platform and/or tenant configuration
  • [PP-2990] - make punchplatform-getconf ready to be used by end user
  • [PP-2998] - standardization of PML examples in standalone
  • [PP-3000] - do not use the Spark REST port
  • [PP-3003] - PML Elasticsearch output node can generate wrong data


  • [PP-2961] - review tests and complete the FileSpouts codecs
  • [PP-2951] - Code Quality - SonarQube integration
  • [PP-3013] - Test Campaign 5.1.2