CloudWatch Logs send log data to a Kinesis data stream that a single consumer Lambda reads and writes to S3. The team observes high latency for some log processing and ingestion. Which actions will reduce that latency? (Choose three.)

Enhanced fan-out allows consumers to receive data from Kinesis Data Streams with dedicated throughput, reducing contention and latency compared to shared throughput. Increasing the ParallelizationFactor for the Lambda event source mapping allows a single Kinesis shard to be processed by multiple concurrent Lambda invocations, improving processing speed. Increasing the number of shards in the Kinesis data stream increases the stream's overall throughput capacity, allowing more data to be ingested and processed in parallel. Setting reserved concurrency for the Lambda function ensures that the function always has a minimum number of concurrent executions available, which helps prevent cold starts and throttling but doesn't directly address high latency from the Kinesis stream itself. Increasing the Kinesis event mapping batch size can increase latency because the Lambda function waits longer to accumulate a larger batch before processing. Disabling ReportBatchItemFailures changes error handling but does not reduce latency.