Skip to main content
Version: v2.3.4

GenGrid Troubleshooting

Overview​

The AlphaSense Enterprise Insight platform provides GenGrid, a grid and table-style Generative AI output capability that enables users to retrieve and display structured AI-generated responses across rows and columns. GenGrid integrates with multiple backend services including retrieval, reranking, citation, and entitlement systems to ground results in relevant document context.

This document serves to outline common failure scenarios related to GenGrid and provide troubleshooting steps for resolution.

Failure Scenarios​

1. Grid Generation Fails or Partially Completes​

Triage:​

The user initiates a GenGrid request and the grid either fails entirely or only partially populates. This is typically caused by unhandled exceptions, memory pressure, or a bad configuration following a recent deployment in the GenGrid execution service.

Troubleshooting:​

Verify that the following services are in Running state on Kubernetes:

kubectl get pods -n research | grep -E "knowledge-workflows-manager|ask-in-doc-gengrid"

Check service logs for errors:

kubectl logs -n research deployment/knowledge-workflows-manager --tail=400

If the service is in an error state, restart it:

kubectl rollout restart deployment/knowledge-workflows-manager -n research
kubectl rollout restart deployment/knowledge-workflows-manager-document-consumer -n research
kubectl rollout restart deployment/knowledge-workflows-manager-queue-consumer -n research
kubectl rollout restart deployment/ask-in-doc-gengrid -n research

2. Results Are Empty or Inconsistent​

Triage:​

The user submits a GenGrid request and the grid completes but returns empty cells, inconsistent data, or results that differ across refreshes. This typically indicates retrieval or reranking degradation where insufficient relevant context is being fetched to ground the response.

Troubleshooting:​

Verify that the following downstream services are in Running state on Kubernetes:

kubectl get pods -n research | grep -E "hybrid-search"

Verify that there are no error logs in the retrieval and reranking services.

Confirm that citation services are healthy and returning results.

3. Timeouts for Larger Grids​

Triage:​

The user submits a GenGrid request with a large number of rows or columns and the request times out or takes an unusually long time to complete. Smaller grids from the same user succeed normally.

Troubleshooting:​

Verify that the GenGrid service is not under resource pressure:

kubectl get pods -n research | grep knowledge-workflows-manager

Check logs for timeout or size-related errors.

If the service is saturated, scale replicas:

kubectl scale deployment/knowledge-workflows-manager -n research --replicas=<N>

If scaling is not appropriate, advise the user to retry with a reduced grid size (fewer rows/columns or a narrower time window) to work within request size limits.

4. Works for Some Users but Not Others​

Triage:​

A GenGrid request succeeds for certain users but fails or returns no results for others using the same query parameters. This is typically caused by entitlement or permission mismatches where the affected user does not have access to the documents used for grounding.

Troubleshooting:​

Confirm whether the problem affects one or multiple users, and whether the issue is consistent or intermittent for the affected users.

Verify that the user exists and is active by opening a port-forward to user-service and querying by username:

# Open a port-forward connection to the user-service in the background
kubectl port-forward -n platform svc/user-service 8080:80 &

# Send a request to get user information
curl 'http://localhost:8080/internal/user/?username=user@domain.com'

Confirm the active attribute is set to true in the response:

{
  "id": 000000,
  "firstName": "Foo",
  "lastName": "Bar",
  "companyId": 0001,
  "email": "user@domain.com",
  "username": "user@domain.com",
  "active": true
}

Check entitlement-related errors in logs for the affected user.

Verify that any shared documents used for grounding are accessible to the affected user and that permission settings are not preventing document retrieval.

5. Cannot Add Items to Grid​

Triage:​

The user is unable to add rows or columns to an existing or new GenGrid. The UI may show an error or the add action may silently fail.

Troubleshooting:​

This issue can be complex to diagnose accurately. It is advisable to create a support ticket with AlphaSense Support for assistance. The following steps serve as initial data collection to facilitate more effective support:

Confirm whether the problem affects one or multiple users, and whether the issue occurs consistently or intermittently.

Collect logs from the GenGrid service.

Save the browser session and create a support request to AlphaSense:

  1. Open a web browser
  2. Browse to the AlphaSense Enterprise Insight platform URL
  3. Reproduce the failure by attempting to add an item to the grid
  4. Open the browser's Developer Tools and save the Network tab as HAR

Provide the following when escalating:

  • Example grid request (rows/columns, query/prompt)
  • Affected user and timestamp
  • Logs showing the failing service hop

Validation Steps​

After applying any resolution:

  1. Verify that all affected pods are in Running state with no recent restarts
  2. Submit a representative GenGrid request and confirm the grid populates completely
  3. If the issue was user-specific, impersonate the affected user and confirm the request now succeeds
  4. If a service was restarted or scaled, monitor logs for healthy operation before closing the incident