Configure the maximum number of items processed concurrently by the Arpio DynamoDB replication function
arpio-config:stream:ParallelizationFactor = <number>
The tag's value must be a number that is valid for the ParallelizationFactor property of the event source mapping that connects the tagged DynamoDB table to the Arpio replicator Lambda function.
- DynamoDB Table
You can use this tag to configure a custom parallelization factor for the event source mapping that's used to handle changed item events for the table the configuration tag is applied to. Each time Arpio starts to create a recovery point, it will update the event source mapping's parallelization factor to match the value in this configuration tag. See Lambda event source mapping batching behavior for information about batching and guidance on setting an appropriate value.
Arpio uses a default parallelization factor of 1. This ensures that items are replicated to the recovery region in the order they were created or changed in the primary table, which is desired for most workloads. If the value is set greater than 1, Arpio may replicate changes for different partition key shards in parallel, but items within a shard will still be replicated in-order.
Most tables do not require a custom parallelization factor to be configured, but tables that support very high rates of change may benefit from a larger batch size. Contact Arpio support for assistance in tuning replication settings for high-throughput tables.