Fix connection after resharding Amazon Connect Kinesis

When the Kinesis data stream that serves Verint interaction capture is resharded, Verint can no longer capture interactions from that stream.

Cause

The Verint Amazon Agent Event Stream Adapter always starts the client connection to Kinesis using the latest checkpoint. It does not perform a checkpoint on shard termination. So after resharding, the Amazon Agent Event Stream Adapter reports multiple, continuous errors in the Integration Server log, and Verint cannot capture new interactions.

Procedure

After you call a resharding operation, either splitShard or mergeShards, wait for the stream to become active again.
In Verint Recorder Manager, go to General Setup, under Integration Adapters, click Settings.
Select the Amazon Agent Stream Adapter.
Take note of the application name before changing it. You need this name later when you delete the old DynamoDB table.
Change the application name in any way, and then click Save.
Restart the Amazon Agent Stream Adapter.

A new DynamoDB table is created for the client connection to the Kinesis stream. The new table contains only new shards, and Verint can now capture interactions.
Delete the old DynamoDB table, as described in the AWS documentation.

The name of the AWS DynamoDB is in the format <application name>_ <RIS server name>.<adapter ID on the RIS server>.

What to do next

Repeat this procedure on the backup Integration Service if one exists.