Kendra's post extract Lambda not working on big collection

Hi, I just created Lambda for Kendra to enrich post extract metadata - calling bedrock to get summary from document, category and add to metadata. It's working fine for 1 or couple documents. But when I call it for index that have over 140 documents, it's stop working after 20-30 documents and fail. I know maximum running time for document enrichment is 60 sec for one lambda running, so I don't know, Kendra isn't calling lambda for every single document instead use this 60 sec for all documents ? Has anyone have an answer for that ? Thanks,