HADOOP-19311: [ABFS] Implement Backoff and Read Footer metrics using IOStatistics Class #7122
Conversation
This file contains 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
added 7 commits
September 16, 2024 00:43
This comment was marked as outdated.
This comment was marked as outdated.
This comment was marked as outdated.
This comment was marked as outdated.
This comment was marked as outdated.
This comment was marked as outdated.
This comment was marked as outdated.
This comment was marked as outdated.
This comment was marked as outdated.
This comment was marked as outdated.
This comment was marked as outdated.
This comment was marked as outdated.
This comment was marked as outdated.
This comment was marked as outdated.
This comment was marked as outdated.
This comment was marked as outdated.
This comment was marked as outdated.
This comment was marked as outdated.
This comment was marked as outdated.
This comment was marked as outdated.
This comment was marked as outdated.
This comment was marked as outdated.
============================================================
|
This comment was marked as outdated.
This comment was marked as outdated.
bhattmanish98
changed the title
bhattmanish98
changed the title
anujmodi2021
requested changes
Dec 17, 2024
hadoop-tools/hadoop-azure/src/main/java/org/apache/hadoop/fs/azurebfs/AbfsCountersImpl.java
Show resolved
Hide resolved
hadoop-tools/hadoop-azure/src/main/java/org/apache/hadoop/fs/azurebfs/enums/RetryValue.java
Show resolved
Hide resolved
hadoop-tools/hadoop-azure/src/main/java/org/apache/hadoop/fs/azurebfs/enums/FileType.java
Show resolved
Hide resolved
...ls/hadoop-azure/src/main/java/org/apache/hadoop/fs/azurebfs/services/AbfsBackoffMetrics.java
Show resolved
Hide resolved
| private void getRetryMetrics(StringBuilder metricBuilder) { | ||
| for (RetryValue retryCount : RETRY_LIST) { | ||
| long totalRequests = getMetricValue(TOTAL_REQUESTS, retryCount); | ||
| metricBuilder.append("$RCTSI$_").append(retryCount.getValue()) |
There was a problem hiding this comment.
+1
All these constants are used specifically for Metric work. They qualify to be added into MetricConstants class. It make sesne to retain that class and add all these constants there itself.
...adoop-azure/src/main/java/org/apache/hadoop/fs/azurebfs/enums/AbfsReadFooterMetricsEnum.java
Outdated
Show resolved
Hide resolved
...ls/hadoop-azure/src/main/java/org/apache/hadoop/fs/azurebfs/services/AbfsBackoffMetrics.java
Outdated
Show resolved
Hide resolved
...ls/hadoop-azure/src/main/java/org/apache/hadoop/fs/azurebfs/services/AbfsBackoffMetrics.java
Outdated
Show resolved
Hide resolved
...ls/hadoop-azure/src/main/java/org/apache/hadoop/fs/azurebfs/services/AbfsBackoffMetrics.java
Outdated
Show resolved
Hide resolved
...-tools/hadoop-azure/src/main/java/org/apache/hadoop/fs/azurebfs/statistics/package-info.java
Outdated
Show resolved
Hide resolved
hadoop-tools/hadoop-azure/src/main/java/org/apache/hadoop/fs/azurebfs/AbfsCountersImpl.java
Show resolved
Hide resolved
...ls/hadoop-azure/src/main/java/org/apache/hadoop/fs/azurebfs/services/AbfsBackoffMetrics.java
Show resolved
Hide resolved
...ls/hadoop-azure/src/main/java/org/apache/hadoop/fs/azurebfs/services/AbfsBackoffMetrics.java
Outdated
Show resolved
Hide resolved
...ls/hadoop-azure/src/main/java/org/apache/hadoop/fs/azurebfs/services/AbfsBackoffMetrics.java
Show resolved
Hide resolved
...hadoop-azure/src/main/java/org/apache/hadoop/fs/azurebfs/services/AbfsReadFooterMetrics.java
Show resolved
Hide resolved
This comment was marked as outdated.
This comment was marked as outdated.
anujmodi2021
approved these changes
Dec 18, 2024
hadoop-tools/hadoop-azure/src/main/java/org/apache/hadoop/fs/azurebfs/enums/FileType.java
Outdated
Show resolved
Hide resolved
|
🎊 +1 overall
This message was automatically generated. |
|
@steveloughran @mukund-thakur @saikatroy038 Could you please review the PR? |
saikatroy038
suggested changes
Jan 6, 2025
| */ | ||
| public class AbfsBackoffMetrics extends AbstractAbfsStatisticsSource { | ||
| private static final List<RetryValue> RETRY_LIST = Arrays.asList( | ||
| ONE, TWO, THREE, FOUR, FIVE_FIFTEEN, FIFTEEN_TWENTY_FIVE, TWENTY_FIVE_AND_ABOVE); |
There was a problem hiding this comment.
Arrays.asList(RetryValue.values())
| return Arrays.stream(AbfsBackoffMetricsEnum.values()) | ||
| .filter(backoffMetricsEnum -> backoffMetricsEnum.getStatisticType().equals(type)) | ||
| .flatMap(backoffMetricsEnum -> | ||
| RETRY.equals(backoffMetricsEnum.getType()) |
There was a problem hiding this comment.
call getMetricName in flatMap instead of repeating the whole logic:
flatMap(backoffMetricsEnum -> RETRY_LIST.stream().map(retryCount -> getMetricName(backoffMetricsEnum, retryCount)))
| * @return the constructed metric name | ||
| */ | ||
| private String getMetricName(AbfsBackoffMetricsEnum metric, RetryValue retryValue) { | ||
| if (RETRY.equals(metric.getType())) { |
| * @param retryValue the retry value | ||
| * @return the constructed metric name | ||
| */ | ||
| private String getMetricName(AbfsBackoffMetricsEnum metric, RetryValue retryValue) { |
| */ | ||
| public class AbfsReadFooterMetrics extends AbstractAbfsStatisticsSource { | ||
| private static final String FOOTER_LENGTH = "20"; | ||
| private static final List<FileType> FILE_TYPE_LIST = Arrays.asList(PARQUET, NON_PARQUET); |
There was a problem hiding this comment.
Arrays.asList(FileType.values())
| .filter(readFooterMetricsEnum -> readFooterMetricsEnum.getStatisticType().equals(type)) | ||
| .flatMap(readFooterMetricsEnum -> | ||
| FILE.equals(readFooterMetricsEnum.getType()) | ||
| ? FILE_TYPE_LIST.stream().map(fileType -> fileType + COLON + readFooterMetricsEnum.getName()) |
There was a problem hiding this comment.
Consider creating a helper method to encapsulate the logic fileType + COLON + readFooterMetricsEnum.getName() and call that method in getMetricNames, getCounterMetricValue, getMeanMetricValue, incrementMetricValue, etc., for better maintainability
| int len, | ||
| long contentLength, | ||
| long nextReadPos) { | ||
| synchronized (this) { |
There was a problem hiding this comment.
do we need synchronized block here? readCount is of type AtomicLong and hence threadsafe
There was a problem hiding this comment.
Make sense, removed synchronized block.
| readFooterMetrics = new AbfsReadFooterMetrics(); | ||
| metricsMap.put(filePathIdentifier, readFooterMetrics); | ||
|
|
||
| private final Map<String, FileTypeMetrics> fileTypeMetricsMap = new HashMap<>(); |
There was a problem hiding this comment.
updateMap, updateReadMetrics call fileTypeMetricsMap.computeIfAbsent which is not threadsafe. Change it to ConcurrentHashMap
| * | ||
| * @param fileTypeMetrics The {@link FileTypeMetrics} object containing metrics and read details. | ||
| * @param len The length of the current read operation. | ||
| */ |
There was a problem hiding this comment.
do we need synchronized here?
The call to addMeanStatisticSample does not require synchronization since IOStatisticsStoreImpl.meanStatisticMap is a ConcurrentHashMap and MeanStatistic.addSample is already synchronized.
If synchronization is needed for any other reason, it should be synchronized on fileTypeMetrics instead of this
There was a problem hiding this comment.
Yes, we need synchronization only to update Mean Statistic, and as rightly mentioned it is already Concurrent so we don't need this. Reverted this.
| * @param fileTypeMetrics The {@link FileTypeMetrics} object containing metrics and read details. | ||
| * @param len The length of the current read operation. | ||
| * @param contentLength The total length of the file content. | ||
| */ |
There was a problem hiding this comment.
synchronized on fileTypeMetrics instead on this
There was a problem hiding this comment.
As mentioned above, no need of this, reverted this.
…19311_AbfsMetricsChanges
|
🎊 +1 overall
This message was automatically generated. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
Description of PR
JIRA: https://issues.apache.org/jira/browse/HADOOP-19311
Current Flow: We have implemented metrics collection in ABFS flow. We have created a custom AbfsBackoffMetrics and AbfsReadFooterMetrics class which stores all the metrics on the file system level. Our objective is to move away from the custom class implementation and use IOStatisticsStore to store the metrics which is present in hadoop-common.
Changes Made: This PR contains the changes related to storing metrics related to above mentioned classes in IOStatisticStore which is present in hadoop-common. AbstractAbfsStatisticsSource abstract class is created which is implementing IOStatisticsSource interface. This will store IOStatistics of the child metrics class.
Both AbfsBackoffMetrics and AbfsReadFooterMetrics is inheriting AbstractAbfsStatisticsSource and store the respective metrics in IOStatisticsStore.