google cloud platform - Vertex AI Data Store indexing doesn't start - Stack Overflow

Summary:Vertex AI Data Store indexing hangs on Document ingestion is working in progress. Document par

Summary:

Vertex AI Data Store indexing hangs on Document ingestion is working in progress. Document parsing and indexing will start later. for each document.

Details:

I use Airbyte to put Confluence pages into BigQuery. Then I use Vertex AI Data Store to index these documents - I have ~1300 documents imported into it and I see Document ingestion is working in progress. Document parsing and indexing will start later. next to each document in the status column.

The documents are imported into data store (when I click View document button I see the data), but the status is as described above. When I use Vertex AI search with Grounding, this data store is not giving any meaningful results.

Surprisingly that has worked for me well on another (private) account a month back, but here on another account (belonging to my company's anization) the indexing doesn't work. I have tried creating the data store in various regions, using periodical sync or one-off sync, using different schemas etc. It always ends up like this.

I confirmed that I can perform billable actions btw. When I create Data Store based on Cloud Storage bucket, it works fine, so it seems it's related to BigQuery only.

Screenshots

Today's import:

Import details:

Confirmed I have not hit the Discovery Engine limits:

Summary:

Vertex AI Data Store indexing hangs on Document ingestion is working in progress. Document parsing and indexing will start later. for each document.

Details:

I use Airbyte to put Confluence pages into BigQuery. Then I use Vertex AI Data Store to index these documents - I have ~1300 documents imported into it and I see Document ingestion is working in progress. Document parsing and indexing will start later. next to each document in the status column.

The documents are imported into data store (when I click View document button I see the data), but the status is as described above. When I use Vertex AI search with Grounding, this data store is not giving any meaningful results.

Surprisingly that has worked for me well on another (private) account a month back, but here on another account (belonging to my company's anization) the indexing doesn't work. I have tried creating the data store in various regions, using periodical sync or one-off sync, using different schemas etc. It always ends up like this.

I confirmed that I can perform billable actions btw. When I create Data Store based on Cloud Storage bucket, it works fine, so it seems it's related to BigQuery only.

Screenshots

Today's import:

Import details:

Confirmed I have not hit the Discovery Engine limits:

Share Improve this question edited Mar 23 at 10:29 sharq asked Mar 23 at 10:16 sharqsharq 1571 silver badge4 bronze badges
Add a comment  | 

1 Answer 1

Reset to default 0

Solved!

Terrible DevEx on Google's side on this one.

I needed to create a Vertex AI app that would actually use this data store before it started indexing the data. The fact that I used this data store in chat for grounding and for custom Cloud Run function did not seem to matter to the data store.

Since it was a data store with multiple entities I had to create a Search type of app.

发布者:admin,转转请注明出处:http://www.yc00.com/questions/1744290328a4566998.html

相关推荐

发表回复

评论列表(0条)

  • 暂无评论

联系我们

400-800-8888

在线咨询: QQ交谈

邮件:admin@example.com

工作时间:周一至周五,9:30-18:30,节假日休息

关注微信