Agent Builder Integration by API

[GuodongJiang-co]

Context: I have built an App by Agent builder by console, in "preview" it can search based on the JSONL in data store

Expected: Integrate the app by API integration, prefer Node.js. Then I can trigger API call and get response from the app engine.

I tried to get it work by either node.js or Python

What I tried:

- Config IAM (cloud storage admin,discovery engine viewer, vertex ai admin)

- Used default ADC and I also created a Service account with(cloud storage admin,discovery engine viewer, vertex ai admin)

- Test REST on cloud shell and was able to get expected response -- success

```

curl -X POST -H "Authorization: Bearer $(gcloud auth print-access-token)" \ -H "Content-Type: application/json" \ "https://discoveryengine.googleapis.com/v1alpha/projects/MY_PROJECT_ID/locations/global/collections/d..." \ -d '{"query":"bank card","pageSize":10,"queryExpansionSpec":{"condition":"AUTO"},"spellCorrectionSpec":{"mode":"AUTO"}}'

```

- Follow

https://cloud.google.com/generative-ai-app-builder/docs/preview-search-results#genappbuilder_search_... code sample. No code updates but populate with my project_id,location, engine_id,api_key and run code from local machine --> GOT 503 error

Need help:
1. how to solve 503 error
2. https://cloud.google.com/generative-ai-app-builder/docs/preview-search-results#genappbuilder_search-... the config doesn't have `engine_id` but needs `dataStoreId`. are these 2 ways (engine OR datastore) similar? I have the app created by agent builder and tuned the model, will the response look similar when I use engine OR datastoreid in config?

[MarvinLlamas]

Hi @GuodongJiang-co,

Welcome to Google Cloud Community!

The 503 error indicates that the server is temporarily unable to handle your project’s request.

Here are some potential ways to address your issue:

• Retry logic: You may integrate a retry mechanism in your code to manage transient errors. This approach can alleviate issues caused by network disruptions or temporary server overloads.
• Examine Logs: You can use Cloud Monitoring to inspect any errors or logs that could provide more insight into the cause of the issue.
• Review IAM permission: Ensure that your service account holds the correct IAM permissions. Make sure it has the required role for accessing the Discovery Engine and other resources.

For engineId and dataStoreId, both are used in configuration and similar in terms of data return, but they serve different purposes.The search engine’s response will depend on the data it pulls from the data store. Therefore, using the correct ID is crucial for getting the desired results.

To give you more insight into the difference between engineId and dataStoreId:

• engineId: Primarily used to identify the Discovery Engine itself. It plays a crucial role in the configuration process, whether querying or setting up the engine.
• dataStoreId: It refers to the ID of the data store within your project to which your app is connected. The data store holds the documents or records that your search engine queries.

I hope the above information is helpful.