snleee commented on code in PR #9249:
URL: https://github.com/apache/pinot/pull/9249#discussion_r950487153


##########
pinot-core/src/main/java/org/apache/pinot/core/data/manager/realtime/RealtimeConsumptionRateManager.java:
##########
@@ -131,4 +144,22 @@ public void throttle(int numMsgs) {
       return _rate;
     }
   }
+
+  @VisibleForTesting
+  @FunctionalInterface
+  interface PartitionCountFetcher {
+    Integer fetch(StreamConfig streamConfig);
+  }
+
+  @VisibleForTesting
+  static final PartitionCountFetcher DEFAULT_PARTITION_COUNT_FETCHER = 
streamConfig -> {
+    String clientId = streamConfig.getTopicName() + 
"-consumption.rate.manager";
+    StreamConsumerFactory factory = 
StreamConsumerFactoryProvider.create(streamConfig);
+    try (StreamMetadataProvider streamMetadataProvider = 
factory.createStreamMetadataProvider(clientId)) {
+      return 
streamMetadataProvider.fetchPartitionCount(/*maxWaitTimeMs*/10_000);
+    } catch (Exception e) {
+      LOGGER.warn("Error fetching metadata for topic " + 
streamConfig.getTopicName(), e);

Review Comment:
   Do we have any metrics for Kafka connection failure on the server side 
consuming segment or the controller side?



##########
pinot-core/src/main/java/org/apache/pinot/core/data/manager/realtime/RealtimeConsumptionRateManager.java:
##########
@@ -85,17 +88,27 @@ public ConsumptionRateLimiter 
createRateLimiter(StreamConfig streamConfig, Strin
     return new RateLimiterImpl(partitionRateLimit);
   }
 
-  private static LoadingCache<StreamConfig, Integer> buildCache() {
-    return 
CacheBuilder.newBuilder().expireAfterWrite(CACHE_ENTRY_EXPIRATION_TIME_IN_MINUTES,
 TimeUnit.MINUTES)
+  @VisibleForTesting
+  static LoadingCache<StreamConfig, Integer> buildCache(PartitionCountFetcher 
partitionCountFetcher,
+      long duration, TimeUnit unit) {
+    return CacheBuilder.newBuilder().refreshAfterWrite(duration, unit)
         .build(new CacheLoader<StreamConfig, Integer>() {
           @Override
-          public Integer load(StreamConfig streamConfig)
+          public Integer load(StreamConfig key)

Review Comment:
   Why don't we use the table name as the key instead of the entire 
`streamConfig`?



-- 
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: commits-unsubscr...@pinot.apache.org

For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscr...@pinot.apache.org
For additional commands, e-mail: commits-h...@pinot.apache.org

Reply via email to