This scenario describes vertical scaling, as defined in the Apache Kafka Streams documentation and general distributed systems terminology. Vertical scaling refers to adding more resources (CPU, memory, or disk) to a single machine to increase processing capacity.

In this example, the Kafka Streams application is moved from a server with four CPU threads to one with ten CPU threads, increasing the available compute resources without increasing the number of application instances. Kafka Streams can take advantage of additional CPU threads through its num.stream.threads configuration, allowing more partitions to be processed in parallel within the same application instance.

Horizontal scaling (also known as scaling out) would involve adding more application instances across multiple servers, not upgrading a single server. Option D is therefore incorrect, and Option C is not a recognized scaling term.

Thus, this example clearly illustrates vertical scaling, where performance is improved by increasing the capacity of a single node.

The Apache Kafka Streams documentation clearly distinguishes between the Kafka Streams DSL and the Processor API. While the DSL is designed for common stream processing operations such as filtering, mapping, joining, and aggregations, it does not support fully custom processing logic.

For use cases that require custom transformations, fine-grained control over record processing, access to headers, timestamps, state stores, and punctuation, Kafka Streams provides the Processor API. This API allows developers to implement custom Processor, Transformer, or ValueTransformer classes and explicitly define the processing topology.

Option A (TopologyTestDriver) is a testing utility, not a development API. Option C (DSL) is higher-level and not suitable for advanced custom logic. Option D does not exist in Kafka.

Therefore, the correct and officially supported approach for defining custom processors in Kafka Streams is to use the Processor API.

=========================

According to the official Apache Kafka Streams documentation, session windows are specifically designed to model periods of activity separated by gaps of inactivity. A session window groups events that occur close together in time and closes when no new events arrive within a defined inactivity gap.

In this use case, tracking user shopping cart behavior requires detecting when a user becomes inactive (for example, stops adding or removing items). Session windows naturally represent this pattern by creating a new window when activity resumes after a period of silence. This makes them ideal for modeling user sessions, cart abandonment detection, and inactivity analysis.

Hopping and tumbling windows (Options B and C) are fixed-time windows and are not well-suited for detecting inactivity, as they segment data based on clock time rather than behavior. Sliding windows are primarily used for continuous joins and do not explicitly model inactivity gaps.

Therefore, session windows are the correct and officially recommended choice for identifying periods of user inactivity in Kafka Streams applications.

=========================

According to the Apache Kafka consumer and scalability documentation, parallelism in message processing is primarily determined by the number of partitions and the number of consumers within a consumer group. Each partition can be assigned to only one consumer at a time. In this scenario, the topic already has four partitions, but only two consumers are processing them, meaning only two partitions are actively consumed in parallel.

Since monitoring indicates that message processing is CPU-bound, the bottleneck is not Kafka I/O but application-side processing. Adding more consumers (up to the number of partitions) allows Kafka to rebalance the partitions and distribute them across more consumer instances, increasing CPU parallelism and reducing lag.

Option A (adding more partitions) can increase parallelism but is more intrusive and may require repartitioning considerations for keys and downstream systems. Option B increases batch size, which can worsen CPU pressure. Option D may reduce per-poll workload but does not increase parallel processing capacity.

Therefore, the correct and recommended action per Kafka documentation is to add more consumers to the consumer group.

=========================

With acks=all, the leader waits for min.insync.replicas to acknowledge the message. Since min.insync.replicas=3, Kafka will only commit the message once 3 brokers (leader + 2 followers) confirm they have the message.

From Kafka Documentation > Acks and Durability:

“If acks=all is specified, the producer will wait until the full set of in-sync replicas has acknowledged the record. The minimum number of in-sync replicas is controlled by min.insync.replicas.”

Even though the replication factor is 4, only 3 acknowledgments are needed, as defined by min.insync.replicas.

According to the official Apache Kafka consumer documentation, the auto.offset.reset configuration is only applied when a consumer has no valid committed offset for a partition. In this scenario, the consumer last committed offset 217, but the topic’s earliest available offset is now 318. This means the previously committed offset is out of range, most likely due to topic retention cleanup.

When a committed offset is out of range, Kafka treats it as invalid and applies the auto.offset.reset policy. Since the consumer is configured with auto.offset.reset=latest, it will start reading from the end of the log, not from the earliest available offset.

Kafka defines the “latest” offset as the log end offset, which is the offset of the next message to be written. Given that the topic’s offsets end at 588, the next offset is 589.

Therefore, the consumer will start reading from offset 589, skipping all existing records in the topic, which aligns with the documented Kafka consumer behavior.

=========================

To isolate logs for a specific connector, you can configure a separate logger and appender in the Connect worker’s log4j.properties file, using the connector’s name as the logging context.

From Kafka Connect Logging Docs:

“Kafka Connect loggers use hierarchical logger names. You can configure per-connector log levels and output files by extending log4j.properties.”

A and C change verbosity but don’t separate logs.

D is false; targeted logging is possible.

A GlobalKTable is a replicated, read-only table available in full on all instances. It's best for reference or lookup datasets such as:

Product catalog

Customer profiles

Warehouse inventory

From Kafka Streams Developer Guide:

“Use GlobalKTable when you need to perform joins using non-partition-aligned reference data that's small enough to replicate.”

Purchases and transactions are high-throughput, time-sensitive streams, not static reference data.

Apache Kafka Connect stores source connector offsets in a Kafka topic defined by the offset.storage.topic configuration. Since this property is explicitly set to connect-offsets, Kafka Connect will commit offsets to that topic, making Option C correct.

The offset.storage.replication.factor controls how many replicas the offsets topic will have. When this value is set to -1, Kafka Connect uses the broker’s default replication factor, as documented in the Kafka Connect worker configuration reference. Therefore, Option A is also correct.

Option B is incorrect because Kafka Connect does not use Kafka’s internal consumer offsets topic (__consumer_offsets). It always uses the configured offsets topic. Option D is incorrect because offset.flush.timeout.ms defines how long Connect will wait for offset commits to complete, not how long it waits before attempting a commit. The commit interval itself is controlled by offset.flush.interval.ms (60 seconds in this case).

Thus, the correct statements that match the expected behavior are A and C.

RetriableException is a subclass of KafkaException, indicating transient issues that might succeed if retried. Examples include:

LeaderNotAvailableException: Happens when metadata is not yet propagated or leader election is in progress.

NotEnoughReplicasException: Happens when the number of in-sync replicas is insufficient. This can be transient if replicas come back online.

From the Apache Kafka Java Client documentation:

"These exceptions (like LeaderNotAvailableException, NotEnoughReplicasException) are transient and the client will retry them automatically depending on retry configuration."

RecordTooLargeException and AuthorizationException are non-retriable as they indicate client-side or permission errors.