mxm commented on code in PR #686:
URL:
https://github.com/apache/flink-kubernetes-operator/pull/686#discussion_r2281625791
##########
flink-autoscaler/src/main/java/org/apache/flink/autoscaler/metrics/ScalingMetrics.java:
##########
@@ -78,6 +88,45 @@ public static void computeDataRateMetrics(
}
}
+ private static Optional<Double> getObservedTpr(
+ Map<FlinkMetric, AggregatedMetric> flinkMetrics,
+ Map<ScalingMetric, Double> scalingMetrics,
+ double numRecordsInPerSecond,
+ Configuration conf) {
+
+ // If there are no incoming records we return infinity to allow scale
down
+ if (numRecordsInPerSecond == 0) {
+ return Optional.of(Double.POSITIVE_INFINITY);
+ }
+
+ // We only measure observed tpr when we are catching up, that is when
the lag is beyond the
+ // configured observe threshold
+ boolean catchingUp =
+ scalingMetrics.getOrDefault(ScalingMetric.LAG, 0.)
+ >=
conf.get(AutoScalerOptions.OBSERVE_TRUE_PROCESSING_RATE_LAG_THRESHOLD)
+ .toSeconds()
+ * numRecordsInPerSecond;
+ if (!catchingUp) {
+ return Optional.empty();
+ }
+
+ double observedTpr =
+ computeObservedTprWithBackpressure(
+ numRecordsInPerSecond,
+
flinkMetrics.get(FlinkMetric.BACKPRESSURE_TIME_PER_SEC).getAvg());
+
+ return Double.isNaN(observedTpr) ? Optional.empty() :
Optional.of(observedTpr);
+ }
+
+ public static double computeObservedTprWithBackpressure(
Review Comment:
True processing rate is the max processing rate in the absence of
backpressure. When we use the task busyness metric, we compute it as
`throughput / (busy_time_ms_per_second / 1000)`. Here, we have only the
throughput which we assume to be the maximum throughput. When there is
backpressure, we need to factor it out to correctly compute the true processing
rate. We do this in both approaches, once by using the busy time metric, once
by using the backpressure metric.
The assumption is that the backpressure comes from downstream and that we
can eliminate it in the next scaling, by resizing downstream operators, and
scale up to the true processing rate.
--
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: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
