vvcephei commented on a change in pull request #9267:
URL: https://github.com/apache/kafka/pull/9267#discussion_r485771434
##########
File path:
streams/src/main/java/org/apache/kafka/streams/processor/internals/StreamThread.java
##########
@@ -706,13 +662,17 @@ void runOnce() {
totalProcessed += processed;
}
+ log.debug("TaskManager#process handled {} records; invoking
TaskManager#punctuate", processed);
+
final int punctuated = taskManager.punctuate();
final long punctuateLatency = advanceNowAndComputeLatency();
totalPunctuateLatency += punctuateLatency;
if (punctuated > 0) {
punctuateSensor.record(punctuateLatency / (double)
punctuated, now);
}
+ log.debug("TaskManager#punctuate executed: {}", punctuated);
Review comment:
Note, this isn't saying that we invoked a single punctuator, but that we
invoked _all_ the punctuators that are runnable right now. It's also a
top-level phase of executing StreamThread, so I don't think we can leave it out
of the debug logs without telling an incomplete story of what's happening. If
we log this (or any of the proposed logs) at trace level, the only consequence
is that users who want to _debug_ StreamThread would have to use "trace" level
instead of "debug" level. This seems to add unnecessary complexity to the
logging.
IMO, it's better to reserve "trace" level for _very_ low-level logs, such as
logging the progress of individual documents through the processors, and use
"debug" level for higher-level summary logs like the ones in this PR.
----------------------------------------------------------------
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
