[client] Implement adaptive fetch rate control for LogScanner #3007
There are no files selected for viewing
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,154 @@ | ||
| /* | ||
| * 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.fluss.client.table.scanner.log; | ||
|
|
||
| import org.apache.fluss.annotation.Internal; | ||
| import org.apache.fluss.annotation.VisibleForTesting; | ||
| import org.apache.fluss.metadata.TableBucket; | ||
|
|
||
| import org.slf4j.Logger; | ||
| import org.slf4j.LoggerFactory; | ||
|
|
||
| import java.util.HashMap; | ||
| import java.util.Map; | ||
|
|
||
| /** | ||
| * Controls the fetch rate for individual buckets based on the amount of data returned in recent | ||
| * fetches. | ||
| * | ||
| * <p>For buckets that consistently return data, they are fetched every round. For buckets that | ||
| * return no data for consecutive fetches, the fetch frequency is progressively reduced using | ||
| * exponential backoff. This is particularly useful for partitioned tables where only the latest | ||
| * partitions have active data, avoiding wasted CPU and network resources on empty partitions. | ||
| * | ||
| * <p>The backoff schedule is: after 1 empty fetch, skip 1 round; after 2 empty fetches, skip 2 | ||
| * rounds; after 3, skip 4 rounds; and so on (powers of 2), up to {@code maxSkipRounds}. Any fetch | ||
| * that returns data immediately resets the backoff to zero. | ||
| * | ||
| * <p>This class is NOT thread-safe. Callers must ensure proper synchronization. | ||
| */ | ||
| @Internal | ||
| class BucketFetchRateController { | ||
|
|
||
| private static final Logger LOG = LoggerFactory.getLogger(BucketFetchRateController.class); | ||
|
|
||
| /** Maximum exponent for the exponential backoff (2^5 = 32). */ | ||
| private static final int MAX_BACKOFF_SHIFT = 5; | ||
|
|
||
| private final int maxSkipRounds; | ||
| private final Map<TableBucket, BucketFetchState> bucketStates; | ||
|
|
||
| BucketFetchRateController(int maxSkipRounds) { | ||
| this.maxSkipRounds = maxSkipRounds; | ||
| this.bucketStates = new HashMap<>(); | ||
| } | ||
|
|
||
| /** | ||
| * Determines whether the given bucket should be included in the current fetch round. | ||
| * | ||
| * <p>If the bucket is in a cool down period (i.e., it has been returning empty results), this | ||
| * method decrements the remaining skip count and returns {@code false}. Otherwise, it returns | ||
| * {@code true} indicating the bucket should be fetched. | ||
| * | ||
| * @param tableBucket the bucket to check | ||
| * @return {@code true} if the bucket should be fetched in this round | ||
| */ | ||
| boolean shouldFetch(TableBucket tableBucket) { | ||
|
Contributor
There was a problem hiding this comment. Is it intentional that a single empty fetch arms a skip? A streaming scanner caught up to HW alternates "batch / empty at HW" every poll, this throttles the empty half even though the bucket is active. If that's by design, worth a comment in the Javadoc |
||
| BucketFetchState state = bucketStates.get(tableBucket); | ||
| if (state == null) { | ||
| return true; | ||
| } | ||
| if (state.remainingSkipRounds > 0) { | ||
| state.remainingSkipRounds--; | ||
| LOG.trace( | ||
| "Adaptive fetch: skipping bucket {} (consecutive empty fetches: {}, " | ||
| + "remaining skip rounds: {})", | ||
| tableBucket, | ||
| state.consecutiveEmptyFetches, | ||
| state.remainingSkipRounds); | ||
| return false; | ||
| } | ||
| return true; | ||
| } | ||
|
|
||
| /** | ||
| * Records the result of a fetch for the given bucket, adjusting future fetch frequency. | ||
| * | ||
| * <p>If the fetch returned data, the backoff is immediately reset to zero. If the fetch | ||
| * returned no data, the consecutive empty count is incremented and a new skip interval is | ||
| * calculated using exponential backoff. | ||
| * | ||
| * @param tableBucket the bucket that was fetched | ||
| * @param hasRecords {@code true} if the fetch returned any records | ||
| */ | ||
| void recordFetchResult(TableBucket tableBucket, boolean hasRecords) { | ||
| BucketFetchState state = | ||
| bucketStates.computeIfAbsent(tableBucket, k -> new BucketFetchState()); | ||
| if (hasRecords) { | ||
| if (state.consecutiveEmptyFetches > 0) { | ||
| LOG.info( | ||
|
|
||
| "Adaptive fetch: bucket {} recovered from cool down " | ||
| + "(was throttled after {} consecutive empty fetches)", | ||
| tableBucket, | ||
| state.consecutiveEmptyFetches); | ||
| } | ||
| state.consecutiveEmptyFetches = 0; | ||
| state.remainingSkipRounds = 0; | ||
| } else { | ||
| state.consecutiveEmptyFetches++; | ||
| int shift = Math.min(state.consecutiveEmptyFetches - 1, MAX_BACKOFF_SHIFT); | ||
| state.remainingSkipRounds = Math.min(1 << shift, maxSkipRounds); | ||
| LOG.info( | ||
|
Contributor
There was a problem hiding this comment. ditto about logging noise |
||
| "Adaptive fetch: bucket {} has no data, entering cool down " | ||
| + "(consecutive empty fetches: {}, will skip {} rounds)", | ||
| tableBucket, | ||
| state.consecutiveEmptyFetches, | ||
| state.remainingSkipRounds); | ||
| } | ||
| } | ||
|
|
||
| /** Removes the tracking state for the given bucket. */ | ||
| void removeBucket(TableBucket tableBucket) { | ||
|
Contributor
There was a problem hiding this comment. Do we actually call it? It seems that without it being called we have a leak |
||
| bucketStates.remove(tableBucket); | ||
| } | ||
|
|
||
| /** Resets all tracking state. */ | ||
| void reset() { | ||
|
|
||
| bucketStates.clear(); | ||
| } | ||
|
|
||
| /** Returns the current number of remaining skip rounds for the given bucket, for testing. */ | ||
| @VisibleForTesting | ||
| int getRemainingSkipRounds(TableBucket tableBucket) { | ||
| BucketFetchState state = bucketStates.get(tableBucket); | ||
| return state == null ? 0 : state.remainingSkipRounds; | ||
| } | ||
|
|
||
| /** Returns the number of consecutive empty fetches for the given bucket, for testing. */ | ||
| @VisibleForTesting | ||
| int getConsecutiveEmptyFetches(TableBucket tableBucket) { | ||
| BucketFetchState state = bucketStates.get(tableBucket); | ||
| return state == null ? 0 : state.consecutiveEmptyFetches; | ||
| } | ||
|
|
||
| /** Per-bucket fetch state tracking consecutive empty fetches and cool down. */ | ||
| private static class BucketFetchState { | ||
| int consecutiveEmptyFetches = 0; | ||
| int remainingSkipRounds = 0; | ||
| } | ||
| } | ||
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Oops, something went wrong.
Oops, something went wrong.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
We also have max_skip_rounds, so with this we have weird situation where user expects config to work, but it's capped at 32.