ableegoldman commented on a change in pull request #8697:
URL: https://github.com/apache/kafka/pull/8697#discussion_r429469651
##########
File path:
streams/src/main/java/org/apache/kafka/streams/processor/internals/metrics/StreamsMetricsImpl.java
##########
@@ -149,6 +154,10 @@ public int hashCode() {
public static final String RATE_DESCRIPTION_PREFIX = "The average number
of ";
public static final String RATE_DESCRIPTION_SUFFIX = " per second";
+ public static final int PERCENTILES_SIZE_IN_BYTES = 1000 * 1000; // 1 MB
+ public static double MAXIMUM_E2E_LATENCY = 100 * 24 * 60 * 60 * 1000d; //
maximum latency is 1000 days
Review comment:
Want to call attention to these...do they seem reasonable? The size is
the bytes per each percentile sensor, so 2 per source or terminal node. The
minimum has to be 0 for the linear bucketing (which I found significantly more
accurate than constant bucketing in my tests).
On the other hand, the maximum is obviously not representative of the
maximum difference between the current and record timestamp. If someone's
processing historical data, it can exceed this. But I figure if you're
processing historical data than the e2e latency isn't really going to be at all
useful anyways, so we may as well set it to something reasonable
----------------------------------------------------------------
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
