This indicates that the logs are having issues getting to Opensearch. The logs go otel-collector -> agent -> gateway -> opni-otel-preprocessor -> opensearch. Would you be able to work your way along that path and check for additional errors? It might be easier to put this in a github issue.
let me double check Ron, but I believe the queue size may be managed internally by our agent kubernetes controller, so it may not be configurable at this time.