mxm commented on code in PR #660: URL: https://github.com/apache/flink-kubernetes-operator/pull/660#discussion_r1309971918
##########
flink-kubernetes-operator-autoscaler/src/main/java/org/apache/flink/kubernetes/operator/autoscaler/AutoscalerFlinkMetrics.java:
##########
@@ -62,54 +61,49 @@ public AutoscalerFlinkMetrics(MetricGroup metricGroup) {
}
public void registerScalingMetrics(
+ Supplier<List<JobVertexID>> jobVerticesSupplier,
Supplier<Map<JobVertexID, Map<ScalingMetric,
EvaluatedScalingMetric>>>
- currentVertexMetrics) {
- currentVertexMetrics
- .get()
- .forEach(
- (jobVertexID, evaluated) -> {
- if (!vertexMetrics.add(jobVertexID)) {
- return;
- }
- LOG.info("Registering scaling metrics for job
vertex {}", jobVertexID);
- var jobVertexMg =
- metricGroup.addGroup(JOB_VERTEX_ID,
jobVertexID.toHexString());
-
- evaluated.forEach(
- (sm, esm) -> {
- var smGroup =
jobVertexMg.addGroup(sm.name());
-
- smGroup.gauge(
- CURRENT,
- () ->
- Optional.ofNullable(
-
currentVertexMetrics.get())
- .map(m ->
m.get(jobVertexID))
- .map(metrics
-> metrics.get(sm))
- .map(
-
EvaluatedScalingMetric
-
::getCurrent)
-
.orElse(Double.NaN));
-
- if (sm.isCalculateAverage()) {
- smGroup.gauge(
- AVERAGE,
- () ->
-
Optional.ofNullable(
-
currentVertexMetrics
-
.get())
- .map(m ->
m.get(jobVertexID))
-
.map(metrics -> metrics.get(sm))
- .map(
-
EvaluatedScalingMetric
-
::getAverage)
-
.orElse(Double.NaN));
- }
- });
- });
+ metricsSupplier) {
+ if (scalingMetricsInitialized) {
+ // It is important that we only initialize these metrics once
because the Flink API does
+ // not support registering counters / gauges multiple times.
+ // The metrics will be updated via the provided metricsSupplier.
+ return;
+ }
+ scalingMetricsInitialized = true;
+
+ LOG.info("Registering scaling metrics");
+ for (JobVertexID jobVertexID : jobVerticesSupplier.get()) {
Review Comment:
Getting the job vertices is an expensive operation. We want to avoid calling
it on every call to this method which will happen for every reconciliation loop.
##########
flink-kubernetes-operator-autoscaler/src/main/java/org/apache/flink/kubernetes/operator/autoscaler/AutoscalerFlinkMetrics.java:
##########
@@ -62,54 +61,49 @@ public AutoscalerFlinkMetrics(MetricGroup metricGroup) {
}
public void registerScalingMetrics(
+ Supplier<List<JobVertexID>> jobVerticesSupplier,
Supplier<Map<JobVertexID, Map<ScalingMetric,
EvaluatedScalingMetric>>>
- currentVertexMetrics) {
- currentVertexMetrics
- .get()
- .forEach(
- (jobVertexID, evaluated) -> {
- if (!vertexMetrics.add(jobVertexID)) {
- return;
- }
- LOG.info("Registering scaling metrics for job
vertex {}", jobVertexID);
- var jobVertexMg =
- metricGroup.addGroup(JOB_VERTEX_ID,
jobVertexID.toHexString());
-
- evaluated.forEach(
- (sm, esm) -> {
- var smGroup =
jobVertexMg.addGroup(sm.name());
-
- smGroup.gauge(
- CURRENT,
- () ->
- Optional.ofNullable(
-
currentVertexMetrics.get())
- .map(m ->
m.get(jobVertexID))
- .map(metrics
-> metrics.get(sm))
- .map(
-
EvaluatedScalingMetric
-
::getCurrent)
-
.orElse(Double.NaN));
-
- if (sm.isCalculateAverage()) {
- smGroup.gauge(
- AVERAGE,
- () ->
-
Optional.ofNullable(
-
currentVertexMetrics
-
.get())
- .map(m ->
m.get(jobVertexID))
-
.map(metrics -> metrics.get(sm))
- .map(
-
EvaluatedScalingMetric
-
::getAverage)
-
.orElse(Double.NaN));
- }
- });
- });
+ metricsSupplier) {
+ if (scalingMetricsInitialized) {
+ // It is important that we only initialize these metrics once
because the Flink API does
+ // not support registering counters / gauges multiple times.
+ // The metrics will be updated via the provided metricsSupplier.
+ return;
+ }
+ scalingMetricsInitialized = true;
+
+ LOG.info("Registering scaling metrics");
Review Comment:
The logger is already scoped by resource id.
--
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]
