What's Changed
[1.2.15] — 2026-04-15
Fixed
- Kinesis
GetRecordsiterator handling — shard iterators are no longer consumed (popped) on use, matching real AWS behavior where iterators remain valid until their 5-minute TTL expires. Previously, callingGetRecordsimmediately invalidated the iterator, causingExpiredIteratorExceptionon client retries. Polling consumers like Apache Camel that retry on transient failures would fail with "Iterator has expired or is invalid". Reported by @markwimpory