cadonna commented on code in PR #12985:
URL: https://github.com/apache/kafka/pull/12985#discussion_r1052198261
##########
streams/src/main/java/org/apache/kafka/streams/processor/internals/RecordCollectorImpl.java:
##########
@@ -159,7 +159,8 @@ public <K, V> void send(final String topic,
final Set<Integer> multicastPartitions =
maybeMulticastPartitions.get();
if (multicastPartitions.isEmpty()) {
// If a record is not to be sent to any partition,
mark it as a dropped record.
- log.debug("Not sending the record with key {} , value
{} to any partition", key, value);
+ log.warn("Skipping record as partitioner returned
empty partitions. "
Review Comment:
I think, we should never log keys and values on any log level. Since the
provider of the Streams service might not own the data, but might be able to
change the log level, having keys and values in the logs seems a security risk
to me.
Regarding, debug vs. warn vs. trace here. In other places, we use warn log
messages when records are dropped since the dropped records metric does not
contain any information about why the record was dropped. However, in this case
I understand -- and correct me if I am wrong -- that the partitioner might drop
records on purpose. So having a warn log message for each record seems not
appropriate. Still it would be great to have some kind of logging. Maybe some
regular summary (e.g., for the first 10 records and then every 100 records and
then on larger intervals?) or indeed logging at trace level.
--
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.
To unsubscribe, e-mail: jira-unsubscr...@kafka.apache.org
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
