search

aws-samples / serverless-pdf-chat

LLM-powered document chat using Amazon Bedrock and AWS Serverless
https://aws.amazon.com/blogs/compute/building-a-serverless-document-chat-with-aws-lambda-and-amazon-bedrock/
MIT No Attribution
245 stars 222 forks source link

Issue with larger files #24

Closed mohsiniqbal368 closed 11 months ago

mohsiniqbal368 commented 11 months ago

It works fine until we keep uploading documents of size in KBs or max 1 to 2 MB with 15 to 20 pages.

I tried uploading a 15MB PDF with roughly 400 pages. It gets uploaded on the S3 bucket but embedding extraction doesn't fire and i am unable to see the document on front end as well after upload.

I also tried deploying with increased timeout along with no spaces in the name of that large pdf file but it still doesn't work.

pbv0 commented 11 months ago

Hi, can you take a look at the Lambda function logs and check what error is coming up? If the document does not show up in the UI, the UploadTrigger function is likely failing.

mohsiniqbal368 commented 11 months ago

All I can see in Cloudwatch for UploadTrigger is 2023-11-20T17:02:37.976Z d7ee6ba3-62c2-45e6-9aac-2d53b6677503 Task timed out after 3.23 seconds

increased the timeout in Lambda configuration tab and its working fine now.

dariushazimi commented 8 months ago

@pbv0 Hi Pascal, Increased the timeout on the lambda func "UploadTrigger function " but a file with about 32MB does not show up on the UI. It is uploaded on the S3 bucket but not listed on UI.