Status

Current state:Accepted Under Discussion

Discussion thread: https://lists.apache.org/thread/yl87h1s484yc09yjo1no46hwpbv0qkwt 

...

Proposed Changes

Controllers

Any errors during metadata processing on the Active Controller cause it to renounce the quorum leadership. These are different than the general Raft elections being triggered due to other reasons like a roll. Repeated elections being caused due to errors in the active controller could point to issues in the metadata log generation handling logic and having visibility into these would be helpful. The ControllerForceRenounceCount metric reflects the number of times a controller node has had to renounce quorum leadership due to an error in the event processing logic.

The ForceRenounceCount metric will be incremented anytime the controller is going to resign as a result of handling exceptions in Event Processing The MetadataErrorCount metric is update for both active and standby controllers. For Active Controllers it is incremented anytime they hit an error in either generating a Metadata log or while applying it to memory. For standby controllers, this metric is incremented when they hit an error in applying the metadata log to memory. This metric will reflect the total count of force resignations errors that a controller (both leader and non-leader) underwent encountered in metadata log processing since the last restart.

https://github.com/apache/kafka/blob/14d2269471141067dc3c45300187f20a0a051777/metadata/src/main/java/org/apache/kafka/controller/QuorumController.java#L409 

 private Throwable handleEventException(String name,
                                           OptionalLong startProcessingTimeNs,
                                           Throwable exception) {
        if (!startProcessingTimeNs.isPresent()) {
        ...
        ...
        renounce();
		//**** Increment ForceRenounceCountMetadataErrorCount
        return new UnknownServerException(exception);
    }

Brokers

The publishermetadata-apply-error-count metric will be incremented by one every time there is an error in publishing a new MetadataImage. This metric will reflect the count of cumulative errors since the broker started up.

...

override def publish(delta: MetadataDelta, newImage: MetadataImage): Unit = {
    val highestOffsetAndEpoch = newImage.highestOffsetAndEpoch()

    try {
      trace(s"Publishing delta $delta with highest offset $highestOffsetAndEpoch")

      // Publish the new metadata image to the metadata cache.
      metadataCache.setImage(newImage)
   	  ...
	  ...
      publishedOffsetAtomic.set(newImage.highestOffsetAndEpoch().offset)
    } catch {
	  //**** Increment publishermetadata-apply-error-count
      case t: Throwable => error(s"Error publishing broker metadata at $highestOffsetAndEpoch", t)	
        throw t
    } finally {
      _firstPublish = false
    }
  }

The listenermetadata-batch-load-error-count metric will be incremented every time there is an error in loading batches and generating MetadataDelta from them. This metric will reflect the count of cumulative errors since the broker started up.

...

class HandleCommitsEvent(reader: BatchReader[ApiMessageAndVersion])
      extends EventQueue.FailureLoggingEvent(log) {
    override def run(): Unit = {
      val results = try {
        val loadResults = loadBatches(_delta, reader, None, None, None)
        ...
        loadResults
      } catch {
		//**** Increment listenermetadata-batch-load-error-count
	  } finally {
        reader.close()
      }

      ...
      _publisher.foreach(publish)
    }
  }

...

  class HandleSnapshotEvent(reader: SnapshotReader[ApiMessageAndVersion])
    extends EventQueue.FailureLoggingEvent(log) {
    override def run(): Unit = {
      try {
        info(s"Loading snapshot ${reader.snapshotId().offset}-${reader.snapshotId().epoch}.")
        _delta = new MetadataDelta(_image) // Discard any previous deltas.
        val loadResults = loadBatches(
       ...
      } catch {
		//**** Increment listener-batchmetadata-load-error-count
	  } finally {
        reader.close()
      }
      _publisher.foreach(publish)
    }
  }

...

Instead of adding the specific metrics, we could have added a more generic MetadataProcessingErrorCount Metric which would be incremented when either of these (and any other similar) or any other similar errors are hit on both Brokers and Controllers. The downside to this approach would be the loss in granularity on what exactly failed on a given node. The specific metrics are more meaningful and give better control over any alarming that might be setup on these metrics.