With SSL, messages will need to be encrypted and decrypted, by being first loaded into the JVM, so you lose the zero copy optimization. See more information herehttps://twitter.com/ijuma/status
/1161303431501324293?s=09

Kafka Connect can run instandalone modeordistributed mode. For simple tasks likereading logs from a file on a single host,standalone modeis recommended.
FromKafka Connect User Guide:
"Standalone mode is useful when running connectors on a single machine (e.g., for development or simple deployments like log collection from a local file)." Distributed mode is preferred for scalability and fault tolerance but overkill for this use case.
Reference:Kafka Connect User Guide > Deployment Modes

When using Kafka Connect with sink connectors (like S3 Sink),each task is assigned one or more topic partitions. Thetotal number of partitions (6)isevenly distributedacross the available tasks (2). Thus, each task will handle3 partitions.
FromKafka Connect Documentation:
"Kafka Connect divides the topic partitions among available tasks. For example, a topic with six partitions and two tasks results in each task handling three partitions." Reference:Kafka Connect Concepts > Tasks and Partitions

Some messages may require multiple retries. If there are more than 1 requests in flight, it may result in messages received out of order. Note an exception to this rule is if you enable the producer settingenable.
idempotence=true which takes care of the out of ordering case on its own. Seehttps://issues.apache.org/jira
/browse/KAFKA-5494

auto.offset.reset=none means that the consumer will crash if the offsets it's recovering from have been deleted from Kafka, which is the case here, as 10 < 45