[jira] [Commented] (BEAM-7219) Spark portable runner: support Spark metrics

["Kyle Weaver (Jira)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/BEAM-7219?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17102995#comment-17102995 ] 

Kyle Weaver commented on BEAM-7219:
-----------------------------------

I'll consider this resolved when test_metrics passes (https://github.com/apache/beam/blob/ad0c5adbb6141ae3dadbb83304c05ed8b44b8d7b/sdks/python/apache_beam/runners/portability/spark_runner_test.py#L115-L117). Right now we still get an error:

$ ./gradlew :sdks:python:test-suites:portable:py2:sparkValidatesRunner -Ptests="test_metrics"
...
ERROR: test_metrics (__main__.SparkRunnerTest)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "apache_beam/runners/portability/fn_api_runner/fn_runner_test.py", line 733, in test_metrics
    'counters']
ValueError: need more than 0 values to unpack


> Spark portable runner: support Spark metrics
> --------------------------------------------
>
>                 Key: BEAM-7219
>                 URL: https://issues.apache.org/jira/browse/BEAM-7219
>             Project: Beam
>          Issue Type: New Feature
>          Components: runner-spark
>            Reporter: Kyle Weaver
>            Assignee: Kyle Weaver
>            Priority: Major
>              Labels: portability-spark
>
> The Spark portable runner already updates and collects metrics from the SDK harness, but it needs to send the collected metrics on to Spark itself, similar to FlinkMetricContainer ([https://github.com/apache/beam/blob/5c31efbae5fa40c024a47a44af3e6bc8a079cc2a/runners/flink/src/main/java/org/apache/beam/runners/flink/metrics/FlinkMetricContainer.java]).



--
This message was sent by Atlassian Jira
(v8.3.4#803005)