enrich-kinesis: Cap recordTtl in KPL

[istreeter]

Until now we set the recordTtl to infinite time when configuring the Kinesis Producer. We thought it was a good idea, because we trusted KPL to retry forever.

However, we found that KPL can go into a death spiral from which it never recovers. I think this happens because KPL retries very aggressively, and it can overwhelm the network, or the Kinesis service. Worst of all, the KPL does not produce any logs while this is happening, so it looks like the app has frozen. If this happens, it would be much better for the KPL to stop retrying, and give up and let the application crash. Therefore, we should cap recordTtl to a finite value.