Skip to content
By using AWS re:Post, you agree to the AWS re:Post Terms of Use
AWS re:Postre:Post
About re:Post

Create a Business Knowledge Assistant quest — unable to complete despite correct setup

0

bug Hi everyone, I’ve finished almost all the quests in "Generative AI Practitioner," but I’m stuck on "Create a Business Knowledge Assistant." I’ve tried everything: changing the CORS policy for the buckets, allowing public access, creating new data sources, reuploading files with different names, and restarting the lab, but nothing worked. The quest requires uploading two files (one .zip and one .pdf) into two different buckets (sales-data and sales-reports). The buckets are prepared for this and an OpenSearch Serverless domain is also created. At this stage the quest asks to create a knowledge base using data sources from the S3 buckets mentioned and to sync them. I’m pretty sure I did everything correctly and I tested the knowledge base using the testing feature with the Nova Lite model; it returned a correct response based on the PDF (the object from the DIY section). Does anyone have any idea how to resolve this?

  • Sambit Sahoo
    3 months ago

    Hi Erick, I am also stuck with this exact same issue. How did you resolve this - can you guide please? Did you wait for a long time and verified again to make it work? Or did you login freshly again and it worked?

Topics
Training and CertificationMachine Learning & AIGenerative AI on AWS
Tags
Knowledge Bases for Amazon BedrockTraining and CertificationAmazon Bedrock
Language
English
asked 3 months ago166 views
2 Answers
1
Accepted Answer

I didn't change my approach and I was able to complete the lab successfully, apparently the problem is fixed :)

answered 3 months ago
  • Sambit Sahoo
    3 months ago

    Hi Erick, I am also stuck with this exact same issue. How did you resolve this - can you guide please? Did you wait for a long time and verified again to make it work? Or did you login freshly again and it worked?

  • Erick Felipe Borges
    3 months ago

    Hi Sambit, I didn’t change my approach. I had already tried everything: changed the buckets’ CORS policy, allowed public access, created new data sources, reuploaded files with different names, and restarted the lab; none of that fixed it. I didn’t repeat those troubleshooting steps; I simply logged in the next day, ran the usual steps again, and it worked. One small detail I noticed while adding the second datasource: on the Review step the knowledge base name shown as the target for that datasource was different from the one I had created, so I canceled and tried again. On the second attempt the Review step showed the knowledge base name empty (which is the correct behavior).

  • Sambit Sahoo
    3 months ago

    Thanks, Erick. It worked for me, but I'm still not sure what the issue was. I was trying on edge, this time I tried on Chrome. Also, I was initially indexing only the sales-data first and then adding second S3 sales-reports and then syncing it. This time I uploaded files in both S3 buckets first, and while creating the collection, applied both syncs. Still not sure why this worked. Wasted 1 complete day for this, definitely some issue with the platform.

0

This appears to be a known issue with AWS Cloud Quest labs, particularly in the Generative AI Practitioner path. Several users have encountered similar problems where they correctly follow all the steps but still can't complete the lab.

For the "Create a Business Knowledge Assistant" quest, even though you've correctly:

  • Uploaded the required files to the appropriate S3 buckets (sales-data and sales-reports)
  • Set up the CORS policy and bucket permissions
  • Created the knowledge base with the proper data sources
  • Successfully tested the knowledge base with the Nova Lite model

The lab validation system may not be recognizing your completion. This could be due to:

  1. Timing issues - sometimes the lab validation checks before the knowledge base has fully processed the data
  2. Specific naming conventions that must be followed exactly
  3. Configuration settings that need to match precisely what the validation is checking for

Some potential solutions to try:

  • Restart the lab completely (not just repeating the steps) as this might reset any underlying configuration issues
  • Double-check that all fields in your knowledge base setup match exactly what's specified in the instructions, including case sensitivity
  • Ensure that the vector database configuration is using Amazon OpenSearch Serverless as specified
  • Try using simple names without special characters for any fields where you provide custom names
  • Wait a bit longer after creating the knowledge base before attempting to complete the lab (to ensure all processing is finished)

If you continue to experience issues despite following all steps correctly, this may be a bug in the Cloud Quest environment that requires AWS support attention. In some cases, lab environments have been known to be fixed after issues are reported.
Sources
Unable to complete AWS Cloud Quest Generative AI Architect task due to SageMaker Studio domain requirement | AWS re:Post
Cloud Quest Gen. AI Architect Task: Issue with "Create Trustworthy LLMs with a Vector Database" Lab | AWS re:Post
Agentic AI Assistant for SAP with AWS Generative AI | AWS for SAP

answered 3 months ago

You are not logged in. Log in to post an answer.

A good answer clearly answers the question and provides constructive feedback and encourages professional growth in the question asker.

Relevant content