Per say, It’s a CloudWatch Log Forwarding (via Kinesis stream with a Lambda hitting ActiveGates – i.e. standard pattern) that the app is spitting 2 different lines of text with exact replica including the timestamp to the nanosec. Regardless, the issue is with ‘Advanced Query’ not able/enabled to identify the uniqueness or factoring for uniqueness ‘validation’.
i.e. Having a duplicate entries in the log is a commonality in log streams, regardless of the ingestion mode.
i.e. perhaps, fix this issue at the 'end of the hose' than 'at the pipe'