shanthoosh commented on a change in pull request #978: SAMZA-1733 : Making
MetricsSnapshotReporter look for appclass then taskclass
URL: https://github.com/apache/samza/pull/978#discussion_r269649574
##########
File path:
samza-core/src/main/scala/org/apache/samza/metrics/reporter/MetricsSnapshotReporterFactory.scala
##########
@@ -45,10 +45,8 @@ class MetricsSnapshotReporterFactory extends
MetricsReporterFactory with Logging
val jobId = config
.getJobId
- val taskClass = config
- .getTaskClass
- .orElse(Option(new ApplicationConfig(config).getAppClass()))
- .getOrElse(throw new SamzaException("No task or app class defined for
config."))
+ val taskClass = Option(new ApplicationConfig(config).getAppClass())
Review comment:
1. I think when this gets used with low-level
API(`StreamTask`/`AsyncStreamTask`) jobs, then `app.class` won't be defined.
2. Would be better to log a message at warn-level that we're falling back to
`task.class` when we don't find `app.class` in config.
3. Within LI, there's yet another configuration `task.subtask.class` where
users specify their task implementation when using offspring-helper. In that
case `task.class` always defaults to a specific samza-predefined-class. Would
be better to think if that would affect any of the functionality here.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
With regards,
Apache Git Services
