You are viewing a plain text version of this content. The canonical link for it is here.
Posted to jira@kafka.apache.org by "chia7712 (via GitHub)" <gi...@apache.org> on 2026/04/17 16:10:09 UTC
Re: [PR] KAFKA-17430: Move RequestChannel.Metrics/RequestMetrics to server module [kafka]
chia7712 commented on code in PR #17015:
URL: https://github.com/apache/kafka/pull/17015#discussion_r3101677920
##########
server/src/main/java/org/apache/kafka/network/metrics/RequestMetrics.java:
##########
@@ -0,0 +1,218 @@
+/*
+ * 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.kafka.network.metrics;
+
+import org.apache.kafka.common.network.ClientInformation;
+import org.apache.kafka.common.protocol.ApiKeys;
+import org.apache.kafka.common.protocol.Errors;
+import org.apache.kafka.server.metrics.KafkaMetricsGroup;
+
+import com.yammer.metrics.core.Histogram;
+import com.yammer.metrics.core.Meter;
+
+import java.util.Collections;
+import java.util.HashMap;
+import java.util.LinkedHashMap;
+import java.util.Map;
+import java.util.Optional;
+import java.util.concurrent.ConcurrentHashMap;
+import java.util.concurrent.ConcurrentMap;
+import java.util.concurrent.TimeUnit;
+
+public class RequestMetrics {
+
+ public static final String CONSUMER_FETCH_METRIC_NAME = ApiKeys.FETCH.name + "Consumer";
+ public static final String FOLLOW_FETCH_METRIC_NAME = ApiKeys.FETCH.name + "Follower";
+ public static final String VERIFY_PARTITIONS_IN_TXN_METRIC_NAME = ApiKeys.ADD_PARTITIONS_TO_TXN.name + "Verification";
+ public static final String REQUESTS_PER_SEC = "RequestsPerSec";
+ public static final String DEPRECATED_REQUESTS_PER_SEC = "DeprecatedRequestsPerSec";
+ public static final String MESSAGE_CONVERSIONS_TIME_MS = "MessageConversionsTimeMs";
+ public static final String TEMPORARY_MEMORY_BYTES = "TemporaryMemoryBytes";
+
+ private static final String REQUEST_QUEUE_TIME_MS = "RequestQueueTimeMs";
+ private static final String LOCAL_TIME_MS = "LocalTimeMs";
+ private static final String REMOTE_TIME_MS = "RemoteTimeMs";
+ private static final String THROTTLE_TIME_MS = "ThrottleTimeMs";
+ private static final String RESPONSE_QUEUE_TIME_MS = "ResponseQueueTimeMs";
+ private static final String RESPONSE_SEND_TIME_MS = "ResponseSendTimeMs";
+ private static final String TOTAL_TIME_MS = "TotalTimeMs";
+ private static final String REQUEST_BYTES = "RequestBytes";
+ private static final String ERRORS_PER_SEC = "ErrorsPerSec";
+
+ // time a request spent in a request queue
+ public final Histogram requestQueueTimeHist;
+ // time a request takes to be processed at the local broker
+ public final Histogram localTimeHist;
+ // time a request takes to wait on remote brokers (currently only relevant to fetch and produce requests)
+ public final Histogram remoteTimeHist;
+ // time a request is throttled, not part of the request processing time (throttling is done at the client level
+ // for clients that support KIP-219 and by muting the channel for the rest)
+ public final Histogram throttleTimeHist;
+ // time a response spent in a response queue
+ public final Histogram responseQueueTimeHist;
+ // time to send the response to the requester
+ public final Histogram responseSendTimeHist;
+ public final Histogram totalTimeHist;
+ // request size in bytes
+ public final Histogram requestBytesHist;
+ // time for message conversions (only relevant to fetch and produce requests)
+ public final Optional<Histogram> messageConversionsTimeHist;
+ // Temporary memory allocated for processing request (only populated for fetch and produce requests)
+ // This shows the memory allocated for compression/conversions excluding the actual request size
+ public final Optional<Histogram> tempMemoryBytesHist;
+
+ // For compatibility - this metrics group was previously defined within
+ // a Scala class named `kafka.network.RequestMetrics`
+ private final KafkaMetricsGroup metricsGroup = new KafkaMetricsGroup("kafka.network", "RequestMetrics");
+ private final String name;
+ private final Map<String, String> tags;
+ private final ConcurrentMap<Short, Meter> requestRateInternal = new ConcurrentHashMap<>();
+ private final ConcurrentMap<DeprecatedRequestRateKey, Meter> deprecatedRequestRateInternal = new ConcurrentHashMap<>();
+ private final Map<Errors, ErrorMeter> errorMeters = new HashMap<>();
+
+ public RequestMetrics(String name) {
+ this.name = name;
+ tags = Collections.singletonMap("request", name);
+ // time a request spent in a request queue
+ requestQueueTimeHist = metricsGroup.newHistogram(REQUEST_QUEUE_TIME_MS, true, tags);
+ // time a request takes to be processed at the local broker
+ localTimeHist = metricsGroup.newHistogram(LOCAL_TIME_MS, true, tags);
+ // time a request takes to wait on remote brokers (currently only relevant to fetch and produce requests)
+ remoteTimeHist = metricsGroup.newHistogram(REMOTE_TIME_MS, true, tags);
+ // time a request is throttled, not part of the request processing time (throttling is done at the client level
+ // for clients that support KIP-219 and by muting the channel for the rest)
+ throttleTimeHist = metricsGroup.newHistogram(THROTTLE_TIME_MS, true, tags);
+ // time a response spent in a response queue
+ responseQueueTimeHist = metricsGroup.newHistogram(RESPONSE_QUEUE_TIME_MS, true, tags);
+ // time to send the response to the requester
+ responseSendTimeHist = metricsGroup.newHistogram(RESPONSE_SEND_TIME_MS, true, tags);
+ totalTimeHist = metricsGroup.newHistogram(TOTAL_TIME_MS, true, tags);
+ // request size in bytes
+ requestBytesHist = metricsGroup.newHistogram(REQUEST_BYTES, true, tags);
+ // time for message conversions (only relevant to fetch and produce requests)
+ messageConversionsTimeHist = isFetchOrProduce(name)
+ ? Optional.of(metricsGroup.newHistogram(MESSAGE_CONVERSIONS_TIME_MS, true, tags))
+ : Optional.empty();
+ // Temporary memory allocated for processing request (only populated for fetch and produce requests)
+ // This shows the memory allocated for compression/conversions excluding the actual request size
+ tempMemoryBytesHist = isFetchOrProduce(name)
+ ? Optional.of(metricsGroup.newHistogram(TEMPORARY_MEMORY_BYTES, true, tags))
+ : Optional.empty();
+ for (Errors error : Errors.values()) {
+ errorMeters.put(error, new ErrorMeter(name, error));
+ }
+ }
+
+ public Meter requestRate(short version) {
+ return requestRateInternal.computeIfAbsent(version, v -> metricsGroup.newMeter(REQUESTS_PER_SEC, "requests", TimeUnit.SECONDS, tagsWithVersion(v)));
+ }
+
+ public Optional<Meter> deprecatedRequestRate(ApiKeys apiKey, short version, ClientInformation clientInformation) {
+ if (apiKey.isVersionDeprecated(version)) {
+ return Optional.of(deprecatedRequestRateInternal.computeIfAbsent(new DeprecatedRequestRateKey(version, clientInformation),
+ k -> metricsGroup.newMeter(DEPRECATED_REQUESTS_PER_SEC, "requests", TimeUnit.SECONDS, tagsWithVersionAndClientInfo(version, clientInformation))));
+ } else {
+ return Optional.empty();
+ }
+ }
+
+ public void markErrorMeter(Errors error, int count) {
+ errorMeters.get(error).getOrCreateMeter().mark(count);
+ }
+
+ private Map<String, String> tagsWithVersion(short version) {
+ Map<String, String> nameAndVersionTags = new LinkedHashMap<>((int) Math.ceil((tags.size() + 1) / 0.75)); // take load factor into account
+ nameAndVersionTags.putAll(tags);
+ nameAndVersionTags.put("version", String.valueOf(version));
+ return nameAndVersionTags;
+ }
+
+ private Map<String, String> tagsWithVersionAndClientInfo(short version, ClientInformation clientInformation) {
+ Map<String, String> extendedTags = new LinkedHashMap<>((int) Math.ceil((tags.size() + 3) / 0.75)); // take load factor into account
+ extendedTags.putAll(tags);
+ extendedTags.put("version", String.valueOf(version));
+ extendedTags.put("clientSoftwareName", clientInformation.softwareName());
+ extendedTags.put("clientSoftwareVersion", clientInformation.softwareVersion());
+ return extendedTags;
+ }
+
+ void removeMetrics() {
+ for (short version : requestRateInternal.keySet()) {
+ metricsGroup.removeMetric(REQUESTS_PER_SEC, tagsWithVersion(version));
+ }
+ for (DeprecatedRequestRateKey key : deprecatedRequestRateInternal.keySet()) {
+ metricsGroup.removeMetric(DEPRECATED_REQUESTS_PER_SEC, tagsWithVersionAndClientInfo(key.version, key.clientInformation));
+ }
+ metricsGroup.removeMetric(REQUEST_QUEUE_TIME_MS, tags);
+ metricsGroup.removeMetric(LOCAL_TIME_MS, tags);
+ metricsGroup.removeMetric(REMOTE_TIME_MS, tags);
+ metricsGroup.removeMetric(REQUESTS_PER_SEC, tags);
+ metricsGroup.removeMetric(THROTTLE_TIME_MS, tags);
+ metricsGroup.removeMetric(RESPONSE_QUEUE_TIME_MS, tags);
+ metricsGroup.removeMetric(TOTAL_TIME_MS, tags);
+ metricsGroup.removeMetric(RESPONSE_SEND_TIME_MS, tags);
+ metricsGroup.removeMetric(REQUEST_BYTES, tags);
+ if (isFetchOrProduce(name)) {
+ metricsGroup.removeMetric(MESSAGE_CONVERSIONS_TIME_MS, tags);
+ metricsGroup.removeMetric(TEMPORARY_MEMORY_BYTES, tags);
+ }
+ for (ErrorMeter errorMeter : errorMeters.values()) {
+ errorMeter.removeMeter();
+ }
+ errorMeters.clear();
+ }
+
+ private boolean isFetchOrProduce(String name) {
+ return ApiKeys.FETCH.name.equals(name) || ApiKeys.PRODUCE.name.equals(name);
+ }
+
+ private class ErrorMeter {
+
+ private final Map<String, String> tags = new LinkedHashMap<>();
+ private volatile Meter meter;
+
+ private ErrorMeter(String name, Errors error) {
+ tags.put("request", name);
+ tags.put("error", error.name());
+ }
+
+ private synchronized Meter getOrCreateMeter() {
Review Comment:
It looks like we missed the double-checked locking here. We'll submit a patch to fix it.
--
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: jira-unsubscribe@kafka.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org