Active Batches piling up with spark streaming with Kafka - scala

I have developed spark streaming(1.6.2) with Kafka in receiver model and running this job with batch size as 15 seconds.
The very first batch is getting a lot of events and processing records very slow. Suddenly job fails and it restarts again. Please see the below screenshot.
It is processing records slowly but not as expected to finish all these batches on time and don't want to see this queue piling up.
How can we control this input size to around 15 to 20k events? I tried enabling spark.streaming.backpressure.enabled but don't see any improvement.
I also implemented Level of Parallelism in Data Receiving as below but still, I didn't see any change in the input size.
val numStreams = 5
val kafkaStreams = (1 to numStreams).map { i => KafkaUtils.createStream(...) }
val unifiedStream = streamingContext.union(kafkaStreams)
unifiedStream.print()
I am using 6 executors and 20 cores.
Overview of my code:
I am reading logs from Kafka and processing them and storing in elasticsearch in every 15 seconds batch interval.
Could you please let me know how can I control input size and improve the performance of the job or how can we make sure that the batches are not pilling up.

Related

Calling Trigger once in Databricks to process Kinesis Stream

I am looking a way to trigger my Databricks notebook once to process Kinesis Stream and using following pattern
import org.apache.spark.sql.streaming.Trigger
// Load your Streaming DataFrame
val sdf = spark.readStream.format("json").schema(my_schema).load("/in/path")
// Perform transformations and then write…
sdf.writeStream.trigger(Trigger.Once).format("delta").start("/out/path")
It looks like it's not possible with AWS Kinesis and that's what Databricks documentation suggest as well. My Question is what else can we do to Achieve that?
As you mentioned in the question the trigger once isn't supported for Kinesis.
But you can achieve what you need by adding into the picture the Kinesis Data Firehose that will write data from Kinesis into S3 bucket (you can select format that you need, like, Parquet, ORC, or just leave in JSON), and then you can point the streaming job to given bucket, and use Trigger.Once for it, as it's a normal streaming source (For efficiency it's better to use Auto Loader that is available on Databricks). Also, to have the costs under the control, you can setup retention policy for your S3 destination to remove or archive files after some period of time, like 1 week or month.
A workaround is to stop after X runs, without trigger. It'll guarantee a fix number of rows per run.
The only issue is that if you have millions of rows waiting in the queue you won't have the guarantee to process all of them
In scala you can add an event listener, in python count the number of batches.
from time import sleep
s = sdf.writeStream.format("delta").start("/out/path")
#by defaut keep spark.sql.streaming.numRecentProgressUpdates=100 in the list. Stop after 10 microbatch
#maxRecordsPerFetch is 10 000 by default, so we will consume a max value of 10x10 000= 100 000 messages per run
while len(s.recentProgress) < 10:
print("Batchs #:"+str(len(s.recentProgress)))
sleep(10)
s.stop()
You can have a more advanced logic counting the number of message processed per batch and stopping when the queue is empty (the throughput should lower once it's all consumed as you'll only get the "real-time" flow, not the history)

Stream kinesis Analytics ETL Flink - skip records before and after a delay

EDITED:
I have a requirement to skip records that are created before 10s and 20s after if a gap in incoming data occurs.
(A gap is said to occur when the event-time1 - event-time2 > 3 seconds)
the resulting data is used to calculate average or median in a timewindow,
Is this possible to be done with Kinesis analytics, Dataflow, flink API, or some solution that works?
If I understand correctly, you want to find the median and average of records that are created between 10 and 20 seconds after a gap of at least 3 seconds.
Using Flink (or Kinesis Analytics, which is a managed Flink service), you could do that with session windows, or with a ProcessFunction. Process functions are more flexible, and are capable of handling pretty much anything you might need. However, in this case, session windows are probably simpler, especially if you are willing to wait until a session ends (i.e., until the next gap) to get the results. You could avoid this delay by implementing a custom window Trigger.
window tutorial
process function tutorial

Flink session window scaling issues on YARN, Kafka

Job use case:
Create groups of events (transactions) that relate to each other, based on event members and event-start-time, from 3 streams, not relying on processing time.
We have a input throughput of arround 20K events/sec.
All events of a transaction are sent to multiple kafka topics (job sources) on the end of a transaction, with the possibility of being late up to minutes.
We want to create event groups that have start times gaps less than a few seconds and are identified by a business key (e.g. country and transaction type).
Transactions can be seconds to hours long, but all events forming a group of transactions should arrive within ~5 minutes after the end of a transaction.
Implementation:
Job consumes 3 data sources, assigns timestamps & watermarks, maps events to a common interface, unions the streams, keys the unioned stream, creates a session window by transaction end time with a lateness, filters groups with huge numbers (bussiness reason), creates sub-groups (based on start-time) within the groups from the session window, and finaly sends the results to another topic.
Issue:
When running the job on a YARN cluster with kafka connectors as I/O, the job is slow, and not scaling well. (~3K events/sec on paralellism 1, to 5K events/sec on parallelism 10)
What I tried:
Playing around with scaling TMs, slots and memory, at max we had arround 10TMs with 2 slot used, each with 32GB memory. (While maintaing the same paralelism on kafka partitions both on input and output)
Playing arround with taskmanager.network.memory.fraction, containerized.heap-cutoff-ratio, taskmanager.memory.fraction
Playing arround with kafka.producer.batch.size, kafka.producer.max.request.size
Setting a fs and then a rocksdb backend and setting taskmanager.memory.off-heap to true, taskmanager.memory.preallocate to true
Turning off checkpointing
Upgrading Flink to 1.9.2, kafka-client to latest used, kafka servers, confluent and cloudera ...
Code:
I recreated a simplified use-case, the code is here: https://github.com/vladimirtomecko/flink-playground and the job looks basically like this:
implicit val (params, config) = PersonTransactionCorrelationConfig.parse(args)
val env: StreamExecutionEnvironment = StreamExecutionEnvironment.getExecutionEnvironment
env.getConfig.setGlobalJobParameters(params)
env.setStreamTimeCharacteristic(TimeCharacteristic.EventTime)
val aTransactions: DataStream[PersonTransactionUniversal] =
env
.addSource(???)
.assignTimestampsAndWatermarks(???)
.flatMap(_.toUniversal: Option[PersonTransactionUniversal])
val bTransactions: DataStream[PersonTransactionUniversal] =
env
.addSource(???)
.assignTimestampsAndWatermarks(???)
.flatMap(_.toUniversal: Option[PersonTransactionUniversal])
val cTransactions: DataStream[PersonTransactionUniversal] =
env
.addSource(???)
.assignTimestampsAndWatermarks(???)
.flatMap(_.toUniversal: Option[PersonTransactionUniversal])
aTransactions
.union(bTransactions)
.union(cTransactions)
.keyBy(new PersonTransactionKeySelector)
.window(EventTimeSessionWindows.withGap(Time.seconds(config.gapEndSeconds)))
.allowedLateness(Time.seconds(config.latenessSeconds))
.aggregate(new PersonTransactionAggregateFunction)
.filter(new PersonTransactionFilterFunction(config.groupMaxSize))
.flatMap(new PersonTransactionFlatMapFunction(config.gapStartSeconds * 1000))
.addSink(new DiscardingSink[PersonTransactionsGroup])
val backend = new RocksDBStateBackend(config.checkpointsDirectory, true)
backend.setPredefinedOptions(PredefinedOptions.SPINNING_DISK_OPTIMIZED)
env.setStateBackend(backend: StateBackend)
env.execute(getClass.getSimpleName)
Question:
Is my implementation incorrect for this use case ?
Is there something I have missed?
Is there some other optimization I can try ?
I have issues finding the bottleneck of this scenario, any tips there ?
Thank you
P.S. First time poster, be kind please.
Edit 1:
Input kafka topics are partitioned on the producer side with the same function used in the keyBy.
Partitions count is equal to or exactly 2 times greater than the parallelism of the flow.
Partitions per topic have a similar amount of events (deviation ~5-10%).
Topics are populated with different amounts of events (A has 10X more events than B, B has 1000X more events than C), but afaik this shouldnt be an issue.

Storm+Kafka not parallelizing as expected

We are having an issue regarding parallelism of tasks inside a single topology. We cannot manage to get a good, fluent processing rate.
We are using Kafka and Storm to build a system with different topologies, where data is processed following a chain of topologies connected using Kafka topics.
We are using Kafka 1.0.0 and Storm 1.2.1.
The load is small in number of messages, about 2000 per day, however each task can take quite some time. One topology in particular can take a variable amount of time to process each task, usually between 1 and 20 minutes. If processed sequentially, the throughput is not enough to process all incoming messages. All topologies and Kafka system are installed in a single machine (16 cores, 16 GB of RAM).
As messages are independent and can be processed in parallel, we are trying to use Storm concurrent capabilities to improve the throughput.
For that the topology has been configured as follows:
4 workers
parallelism hint set to 10
Message size when reading from Kafka large enough to read about 8 tasks in each message.
Kafka topics use replication-factor = 1 and partitions = 10.
With this configuration, we observe the following behavior in this topology.
About 7-8 tasks are read in a single batch from Kafka by the Storm topology (task size is not fixed), max message size of 128 kB.
About 4-5 task are computed concurrently. Work is more-or-less evenly distributed among workers. Some workers take 1 task, others take 2 and process them concurrently.
As tasks are being finished, the remaining tasks start.
A starvation problem is reached when only 1-2 tasks remain to be processed. Other workers wait idle until all tasks are finished.
When all tasks are finished, the message is confirmed and sent to the next topology.
A new batch is read from Kafka and the process starts again.
We have two main issues. First, even with 4 workers and 10 parallelism hint, only 4-5 tasks are started. Second, no more batches are started while there is work pending, even if it is just 1 task.
It is not a problem of not having enough work to do, as we tried inserting 2000 tasks at the beginning, so there is plenty of work to do.
We have tried to increase the parameter "maxSpoutsPending", expecting that the topology would read more batches and queue them at the same time, but it seems they are being pipelined internally, and not processed concurrently.
Topology is created using the following code:
private static StormTopology buildTopologyOD() {
//This is the marker interface BrokerHosts.
BrokerHosts hosts = new ZkHosts(configuration.getProperty(ZKHOSTS));
TridentKafkaConfig tridentConfigCorrelation = new TridentKafkaConfig(hosts, configuration.getProperty(TOPIC_FROM_CORRELATOR_NAME));
tridentConfigCorrelation.scheme = new RawMultiScheme();
tridentConfigCorrelation.fetchSizeBytes = Integer.parseInt(configuration.getProperty(MAX_SIZE_BYTES_CORRELATED_STREAM));
OpaqueTridentKafkaSpout spoutCorrelator = new OpaqueTridentKafkaSpout(tridentConfigCorrelation);
TridentTopology topology = new TridentTopology();
Stream existingObject = topology.newStream("kafka_spout_od1", spoutCorrelator)
.shuffle()
.each(new Fields("bytes"), new ProcessTask(), new Fields(RESULT_FIELD, OBJECT_FIELD))
.parallelismHint(Integer.parseInt(configuration.getProperty(PARALLELISM_HINT)));
//Create a state Factory to produce outputs to kafka topics.
TridentKafkaStateFactory stateFactory = new TridentKafkaStateFactory()
.withProducerProperties(kafkaProperties)
.withKafkaTopicSelector(new ODTopicSelector())
.withTridentTupleToKafkaMapper(new ODTupleToKafkaMapper());
existingObject.partitionPersist(stateFactory, new Fields(RESULT_FIELD, OBJECT_FIELD), new TridentKafkaUpdater(), new Fields(OBJECT_FIELD));
return topology.build();
}
and config created as:
private static Config createConfig(boolean local) {
Config conf = new Config();
conf.setMaxSpoutPending(1); // Also tried 2..6
conf.setNumWorkers(4);
return conf;
}
Is there anything we can do to improve the performance, either by increasing the number of parallel tasks or/and avoiding starvation while finishing to process a batch?
I found an old post on storm-users by Nathan Marz regarding setting parallelism for Trident:
I recommend using the "name" function to name portions of your stream
so that the UI shows you what bolts correspond to what sections.
Trident packs operations into as few bolts as possible. In addition,
it never repartitions your stream unless you've done an operation
that explicitly involves a repartitioning (e.g. shuffle, groupBy,
partitionBy, global aggregation, etc). This property of Trident
ensures that you can control the ordering/semi-ordering of how things
are processed. So in this case, everything before the groupBy has to
have the same parallelism or else Trident would have to repartition
the stream. And since you didn't say you wanted the stream
repartitioned, it can't do that. You can get a different parallelism
for the spout vs. the each's following by introducing a repartitioning
operation, like so:
stream.parallelismHint(1).shuffle().each(…).each(…).parallelismHint(3).groupBy(…);
I think you might want to set parallelismHint for your spout as well as your .each.
Regarding processing multiple batches concurrently, you are right that that is what maxSpoutPending is for in Trident. Try checking in Storm UI that your max spout pending value is actually picked up. Also try enabling debug logging for the MasterBatchCoordinator. You should be able to tell from that logging whether multiple batches are in flight at the same time or not.
When you say that multiple batches are not processed concurrently, do you mean by ProcessTask? Keep in mind that one of the properties of Trident is that state updates are ordered with regard to batches. If you have e.g. maxSpoutPending=3 and batch 1, 2 and 3 in flight, Trident won't emit more batches for processing until batch 1 is written, at which point it'll emit one more. So slow batches can block emitting more, even if 2 and 3 are fully processed, they have to wait for 1 to finish and get written.
If you don't need the batching and ordering behavior of Trident, you could try regular Storm instead.
More of a side note, but you might want to consider migrating off storm-kafka to storm-kafka-client. It's not important for this question, but you won't be able to upgrade to Kafka 2.x without doing it, and it's easier before you get a bunch of state to migrate.

Spark Streaming mapWithState seems to rebuild complete state periodically

I am working on a Scala (2.11) / Spark (1.6.1) streaming project and using mapWithState() to keep track of seen data from previous batches.
The state is distributed in 20 partitions on multiple nodes, created with StateSpec.function(trackStateFunc _).numPartitions(20). In this state we have only a few keys (~100) mapped to Sets with up ~160.000 entries, which grow throughout the application. The entire state is up to 3GB, which can be handled by each node in the cluster. In each batch, some data is added to a state but not deleted until the very end of the process, i.e. ~15 minutes.
While following the application UI, every 10th batch's processing time is very high compared to the other batches. See images:
The yellow fields represent the high processing time.
A more detailed Job view shows that in these batches occur at a certain point, exactly when all 20 partitions are "skipped". Or this is what the UI says.
My understanding of skipped is that each state partition is one possible task which isn't executed, as it doesn't need to be recomputed. However, I don't understand why the amount of skips varies in each Job and why the last Job requires so much processing. The higher processing time occurs regardless of the state's size, it just impacts the duration.
Is this a bug in the mapWithState() functionality or is this intended behaviour? Does the underlying data structure require some kind of reshuffling, does the Set in the state need to copy data? Or is it more likely to be a flaw in my application?
Is this a bug in the mapWithState() functionality or is this intended
behaviour?
This is intended behavior. The spikes you're seeing is because your data is getting checkpointed at the end of that given batch. If you'll notice the time on the longer batches, you'll see that it happens persistently every 100 seconds. That's because the checkpoint time is constant, and is calculated per your batchDuration, which is how often you talk to your data source to read a batch multiplied by some constant, unless you explicitly set the DStream.checkpoint interval.
Here is the relevant piece of code from MapWithStateDStream:
override def initialize(time: Time): Unit = {
if (checkpointDuration == null) {
checkpointDuration = slideDuration * DEFAULT_CHECKPOINT_DURATION_MULTIPLIER
}
super.initialize(time)
}
Where DEFAULT_CHECKPOINT_DURATION_MULTIPLIER is:
private[streaming] object InternalMapWithStateDStream {
private val DEFAULT_CHECKPOINT_DURATION_MULTIPLIER = 10
}
Which lines up exactly with the behavior you're seeing, since your read batch duration is every 10 seconds => 10 * 10 = 100 seconds.
This is normal, and that is the cost of persisting state with Spark. An optimization on your side could be to think how you can minimize the size of the state you have to keep in memory, in order for this serialization to be as quick as possible. Additionaly, make sure that the data is spread out throughout enough executors, so that state is distributed uniformly between all nodes. Also, I hope you've turned on Kryo Serialization instead of the default Java serialization, that can give you a meaningful performance boost.
In addition to the accepted answer, pointing out the price of serialization related to checkpointing, there's another, less known issue which might contribute to the spikey behaviour: eviction of deleted states.
Specifically, 'deleted' or 'timed out' states are not removed immediately from the map, but are marked for deletion and actually removed only in the process of serialization [in Spark 1.6.1, see writeObjectInternal()].
This has two performance implications, which occur only once per 10 batches:
The traversal and deletion process has its price
If you process the stream of timed-out/ deleted events, e.g. persist it to external storage, the associated cost for all 10 batches will be paid only at this point (and not as one might have expected, on each RDD)