[
https://issues.apache.org/jira/browse/FLINK-7781?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16224586#comment-16224586
]
ASF GitHub Bot commented on FLINK-7781:
---------------------------------------
Github user kl0u commented on a diff in the pull request:
https://github.com/apache/flink/pull/4901#discussion_r147650987
--- Diff:
flink-runtime/src/main/java/org/apache/flink/runtime/rest/handler/legacy/metrics/AbstractAggregatingMetricsHandler.java
---
@@ -0,0 +1,223 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one
+ * or more contributor license agreements. See the NOTICE file
+ * distributed with this work for additional information
+ * regarding copyright ownership. The ASF licenses this file
+ * to you under the Apache License, Version 2.0 (the
+ * "License"); you may not use this file except in compliance
+ * with the License. You may obtain a copy of the License at
+ *
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+package org.apache.flink.runtime.rest.handler.legacy.metrics;
+
+import org.apache.flink.runtime.jobmaster.JobManagerGateway;
+import
org.apache.flink.runtime.rest.handler.legacy.AbstractJsonRequestHandler;
+import org.apache.flink.runtime.rest.handler.legacy.JsonFactory;
+import org.apache.flink.util.FlinkException;
+import org.apache.flink.util.Preconditions;
+
+import com.fasterxml.jackson.core.JsonGenerator;
+import org.slf4j.Logger;
+import org.slf4j.LoggerFactory;
+
+import java.io.IOException;
+import java.io.StringWriter;
+import java.util.ArrayList;
+import java.util.Collection;
+import java.util.HashSet;
+import java.util.Iterator;
+import java.util.List;
+import java.util.Map;
+import java.util.Set;
+import java.util.concurrent.CompletableFuture;
+import java.util.concurrent.CompletionException;
+import java.util.concurrent.Executor;
+
+/**
+ * Abstract request handler for querying metrics. Subclasses may either
that returns a list of all available metrics or
+ * the values for a set of metrics for a specific entity, or aggregate
them across several entities.
+ *
+ * <p>If the query parameters do not contain a "get" parameter the list of
all metrics is returned.
+ * {@code [ { "id" : "X" } ] }
+ *
+ * <p>If the query parameters do contain a "get" parameter, a
comma-separated list of metric names is expected as a value.
+ * {@code /metrics?get=X,Y}
+ * The handler will then return a list containing the values of the
requested metrics.
+ * {@code [ { "id" : "X", "value" : "S" }, { "id" : "Y", "value" : "T" } ]
}
+ *
+ * <p>If the query parameters do contain a "get" parameter a
comma-separate list of metric names is expected as a value.
+ * The "agg" query parameter is used to define which aggregates should be
calculated. Available aggregations are
+ * "sum", "max", "min" and "avg". If the parameter is not specified all
aggregations will be returned.
+ * {@code /metrics?get=X,Y&agg=min,max}
+ * The handler will then return an object containing the aggregations for
the requested metrics.
+ * {@code [ { "id" : "X", "min", "1", "max", "2" }, { "id" : "Y", "min",
"4", "max", "10"}]}
+ */
+abstract class AbstractAggregatingMetricsHandler extends
AbstractJsonRequestHandler {
+
+ protected final Logger log = LoggerFactory.getLogger(getClass());
+
+ private static final String PARAMETER_AGGREGATION = "agg";
+
+ private final MetricFetcher fetcher;
+
+ AbstractAggregatingMetricsHandler(Executor executor, MetricFetcher
fetcher) {
+ super(executor);
+ this.fetcher = Preconditions.checkNotNull(fetcher);
+ }
+
+ protected abstract Collection<? extends
MetricStore.ComponentMetricStore> getStores(MetricStore store, Map<String,
String> pathParameters, Map<String, String> queryParameters);
+
+ @Override
+ public CompletableFuture<String> handleJsonRequest(Map<String, String>
pathParams, Map<String, String> queryParams, JobManagerGateway
jobManagerGateway) {
+ return CompletableFuture.supplyAsync(
+ () -> {
+ try {
+ fetcher.update();
+ String requestedMetricsList =
queryParams.get(AbstractMetricsHandler.PARAMETER_METRICS);
+ String aggTypeList =
queryParams.get(PARAMETER_AGGREGATION);
+ MetricStore store =
fetcher.getMetricStore();
+
+ if (requestedMetricsList == null) {
+ Collection<? extends
MetricStore.ComponentMetricStore> stores = getStores(store, pathParams,
queryParams);
+ if (stores == null) {
+ return "[]";
+ }
+ Collection<String> list =
getAvailableMetrics(stores);
+ return
mapMetricListToJson(list);
+ }
+
+ if (requestedMetricsList.isEmpty()) {
+ /*
+ * The WebInterface doesn't
check whether the list of available metrics was empty. This can lead to a
+ * request for which the "get"
parameter is an empty string.
+ */
+ return "[]";
+ }
+
+ String[] requestedMetrics =
requestedMetricsList.split(",");
+
+
List<DoubleAccumulator.DoubleAccumulatorFactory<?>>
requestedAggregationsFactories = new ArrayList<>();
+ // by default we return all aggregations
+ if (aggTypeList == null ||
aggTypeList.isEmpty()) {
+
requestedAggregationsFactories.add(DoubleAccumulator.DoubleMinimumFactory.get());
+
requestedAggregationsFactories.add(DoubleAccumulator.DoubleMaximumFactory.get());
+
requestedAggregationsFactories.add(DoubleAccumulator.DoubleSumFactory.get());
+
requestedAggregationsFactories.add(DoubleAccumulator.DoubleAverageFactory.get());
+ } else {
+ for (String aggregation :
aggTypeList.split(",")) {
+ switch
(aggregation.toLowerCase()) {
+ case
DoubleAccumulator.DoubleMinimum.NAME:
+
requestedAggregationsFactories.add(DoubleAccumulator.DoubleMinimumFactory.get());
+ break;
+ case
DoubleAccumulator.DoubleMaximum.NAME:
+
requestedAggregationsFactories.add(DoubleAccumulator.DoubleMaximumFactory.get());
+ break;
+ case
DoubleAccumulator.DoubleSum.NAME:
+
requestedAggregationsFactories.add(DoubleAccumulator.DoubleSumFactory.get());
+ break;
+ case
DoubleAccumulator.DoubleAverage.NAME:
+
requestedAggregationsFactories.add(DoubleAccumulator.DoubleAverageFactory.get());
+ break;
+ default:
+
log.warn("Invalid aggregation specified: {}", aggregation.toLowerCase());
+ }
+ }
+ }
+ Collection<? extends
MetricStore.ComponentMetricStore> stores = getStores(store, pathParams,
queryParams);
--- End diff --
Shouldn't this check for the `stores` be before populating the
`requestedAggregationsFactories`? If the `stores` is empty, then object
creation and all the work previously done is in vain.
Can this check be factored out before even checking the
`requestedMetricsList`? So that we avoid code repetition. It seems that always
when the `stores==null`, we `return "[]"`.
> Support simple on-demand metrics aggregation
> --------------------------------------------
>
> Key: FLINK-7781
> URL: https://issues.apache.org/jira/browse/FLINK-7781
> Project: Flink
> Issue Type: Improvement
> Components: Metrics, REST
> Affects Versions: 1.4.0
> Reporter: Chesnay Schepler
> Assignee: Chesnay Schepler
> Fix For: 1.4.0
>
>
> We should support aggregations (min, max, avg, sum) of metrics in the REST
> API. This is primarily about aggregating across subtasks, for example the
> number of incoming records across all subtasks.
> This is useful for simple use-cases where a dedicated metrics backend is
> overkill, and will allow us to provide better metrics in the web UI (since we
> can expose these aggregated as well).
> I propose to add a new query parameter "agg=[min,max,avg,sum]". As a start
> this parameter should only be used for task metrics. (This is simply the main
> use-case i have in mind)
> The aggregation should (naturally) only work for numeric metrics.
> We will need a HashSet of metrics that exist for subtasks of a given tasks
> that has to be updated in {{MetricStore#add}}.
> All task metrics are either stored as
> # {{<subtask-index>.<metric>}} or
> # {{<subtask-index>.<operator-name>.<metric>}}.
> If a user sends a request {{get=mymetric,agg=sum}}, only the metrics of the
> first kind are to be considered. Similarly, given a request
> {{get=myoperator.mymetric,agg=sum}} only metrics of the second kind are to be
> considered.
> Ideally, the name of the aggregated metric (i.e. the original name without
> subtask index) is also contained in the list of available metrics.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
