Druid Kafka ingestion with read-your-writes - apache-kafka

I'm learning Druid now. I read that ingestion via Kafka Indexing Service guarantees exactly-once semantics.
However, I have a problem with determining consistency model of Druid. Typically streams are asynchronous, but I want to have read-your-writes semantics in application.
Is there any possibility to check Druid's ingestion status? For example, I send event A and want to check if it was already saved in Druid. If yes, query to Druid should return result with this value.
Maybe there is some other possibility to do real-time ingestion with exactly-once semantics and with read-your-writes?

Druid has separate process for ingestion and reading the data. read-your-writes won't be directly possible, however you can get the success for writes, and than you can make a separate query for reading your write.
check out tranquility server, which gives an http based gateway to write in real-time and it tries to handle exactly once ingestion too.
Though the best approach to ensure exactly once ingestion is to do reindexing by batch ingestion at regular interval depending on your use case.

Related

Stream CDC change with Kafka and Spark still processes it in batches, whereas we wish to process each record

I'm still new in Spark and I want to learn more about it. I want to build and data pipeline architecture with Kafka and Spark.Here is my proposed architecture where PostgreSQL provide data for Kafka. The condition is the PostgreSQL are not empty and I want to catch any CDC change in the database. At the end,I want to grab the Kafka Message and process it in stream with Spark so i can get analysis about what happen at the same time when the CDC event happen.
However, when I try to run an simple stream, it seems Spark receive the data in stream, but process the data in batch, which not my goal. I have see some article that the source of data for this case came from API which we want to monitor, and there's limited case for Database to Database streaming processing. I have done the process before with Kafka to another database, but i need to transform and aggregate the data (I'm not use Confluent and rely on generic Kafka+Debezium+JDBC connectors)
According to my case, is Spark and Kafka can meet the requirement? Thank You
I have designed such pipelines and if you use Structured Streaming KAFKA in continuous or non-continuous mode, you will always get a microbatch. You can process the individual records, so not sure what the issue is.
If you want to process per record, then use the Spring Boot KAFKA setup for consumption of KAFKA messages, that can work in various ways, and fulfill your need. Spring Boor offers various modes of consumption.
Of course Spark Structured Streaming can be done using Scala and has a lot of support obviating extra work elsewhere.
https://medium.com/#contactsunny/simple-apache-kafka-producer-and-consumer-using-spring-boot-41be672f4e2b This article discusses the single message processing approach.

Kafka streams exactly once processing use case

I have use case where i need to read data from topic then batch data(100 records) and write the batch to specific file or external store. I am planning to use processor API for this and batch the data in process method using state store backed by kafka and write to file once the batch size reaches 100 records. Clear the batch from the state store to create fresh new batch.
One more requirements is that we cannot have duplicates in data. This mean same record cannot be in two different batches.
Does streams exactly once fit this use case?? I read in the design that its not recommended if we are batching data and most of the articles around this say that Exactly once works only in the case of consume process and produce pattern.
Kafka Stream's exactly once does only work if you write the result back to Kafka. Because you want to write data to an external system, Kafka cannot provide any help for exactly-once guarantees, because Kafka transactions are not cross-system transactions.
As pointed out #Matthias, Exactly one semantics only work with Kafka streams to Kafka streams type application, integration with an external system is likely to break the semantics. You can read more about it in this article.
I would suggest you use Kafka Consumer API as it will provide the best balance between flexibility and abstraction for your use case. All you need to do is to remove enable.auto.commit=false and manually commit after successfully writing the batch to the external system using consumer.commitSync();
Ensuring exactly once can get a little difficult sometimes depending on your use case. You'll need to make sure that your consumer is idempotent using custom logic. You can consider using external persistent storage to keep to hash (or the key if it is unique) of the messages and check for each message if it is not already processed. You can also use state store for this purpose but I have felt that clearing a state store sometimes becomes a hassle, but it depends a lot on your use case.
You can check out this article if it helps.

Is there a way that i can push historical data into druid over http?

I have an IOT project and want to use Druid as Time Series DBMS. Sometimes the IOT device may lose the network and will re-transfer the historical data and real-time data when reconnecting to the server. I know the Druid can ingest real-time data over http push/pull and historical data over http pull or KIS, but i can't find the document about ingesting historical data over http push.
Is there a way that i can send historical data into druid over http push?
I see a few options here:
Keep pushing historical data to the same kafka topic (or other streaming source) and do a rejection based on message-timestamp inside Druid. This simplifies your application architecture and let druid handle expired events rejection
Use batch ingestion for historical data. You push the historical data to another Kafka topic, run a spark/gobblin/any other index job to get the data to HDFS. Then do a batch ingestion onto Druid. But remember that Druid overwrites any real-time segments with batch segments for the specified windowPeriod. So if the historical data is not complete, you run into data loss. To prevent this, you could always pump real-time data into hadoop as well and do a de-duplication on the HDFS data periodically and ingest into Druid. As you can see this is a complicated architecture, but this can result in minimal data loss.
If I were you, I would simplify and send all data to the same streaming source like Kafka. I would index segments in Druid based on my message's timestamp and not current time (which is the default I believe).
kafka indexing service released recently guarantees exactly once ingestion.
Refer the below link - http://druid.io/docs/latest/development/extensions-core/kafka-ingestion.html
If you still want to ingest over http, you can checkout tranquility server. It has some mechanisms built-in for handling duplicates.

Reliable usage of DirectKafkaAPI

I am pllaned to develop a reliable streamig application based on directkafkaAPI..I will have one producer and another consumer..I wnated to know what is the best approach to achieve the reliability in my consumer?..I can employ two solutions..
Increasing the retention time of messages in Kafka
Using writeahead logs
I am abit confused regarding the usage of writeahead logs in directkafka API as there is no receiver..but in the documentation it indicates..
"Exactly-once semantics: The first approach uses Kafka’s high level API to store consumed offsets in Zookeeper. This is traditionally the way to consume data from Kafka. While this approach (in combination with write ahead logs) can ensure zero data loss (i.e. at-least once semantics), there is a small chance some records may get consumed twice under some failures. "
so I wanted to know what is the best approach..if it suffices to increase the TTL of messages in kafka or I have to also enable write ahead logs..
I guess it would be good practice if I avoid one of the above since the backup data (retentioned messages, checkpoint files) can be lost and then recovery could face failure..
Direct Approach eliminates the duplication of data problem as there is no receiver, and hence no need for Write Ahead Logs. As long as you have sufficient Kafka retention, messages can be recovered from Kafka.
Also, Direct approach by default supports exactly-once message delivery semantics, it does not use Zookeeper. Offsets are tracked by Spark Streaming within its checkpoints.

Spark/Spark Streaming in production without HDFS

I have been developing applications using Spark/Spark-Streaming but so far always used HDFS for file storage. However, I have reached a stage where I am exploring if it can be done (in production, running 24/7) without HDFS. I tried sieving though Spark user group but have not found any concrete answer so far. Note that I do use checkpoints and stateful stream processing using updateStateByKey.
Depending on the streaming(I've been using Kafka), you do not need to use checkpoints etc.
Since spark 1.3 they have implemented a direct approach with so many benefits.
Simplified Parallelism: No need to create multiple input Kafka streams
and union-ing them. With directStream, Spark Streaming will create as
many RDD partitions as there is Kafka partitions to consume, which
will all read data from Kafka in parallel. So there is one-to-one
mapping between Kafka and RDD partitions, which is easier to
understand and tune.
Efficiency: Achieving zero-data loss in the first approach required
the data to be stored in a Write Ahead Log, which further replicated
the data. This is actually inefficient as the data effectively gets
replicated twice - once by Kafka, and a second time by the Write Ahead
Log. This second approach eliminate the problem as there is no
receiver, and hence no need for Write Ahead Logs.
Exactly-once semantics: The first approach uses Kafka’s high level API
to store consumed offsets in Zookeeper. This is traditionally the way
to consume data from Kafka. While this approach (in combination with
write ahead logs) can ensure zero data loss (i.e. at-least once
semantics), there is a small chance some records may get consumed
twice under some failures. This occurs because of inconsistencies
between data reliably received by Spark Streaming and offsets tracked
by Zookeeper. Hence, in this second approach, we use simple Kafka API
that does not use Zookeeper and offsets tracked only by Spark
Streaming within its checkpoints. This eliminates inconsistencies
between Spark Streaming and Zookeeper/Kafka, and so each record is
received by Spark Streaming effectively exactly once despite failures.
If you are using Kafka, you can found out more here:
https://spark.apache.org/docs/1.3.0/streaming-kafka-integration.html
Approach 2.