[FEATURE] Add an ID Field to Search Query Records

[ansjcy]

Is your feature request related to a problem?

To support fetching a specific search query record by ID (e.g., for the current detail page or inspecting a historical slow query in the future), we need to add a unique identifier to each search query record.

What solution would you like?

Add an ID field to search query records.

• Preferred Approach: Use a randomly generated UUID (via UUID.randomUUID() in Java). UUIDs are 128 bits and provide ~3.4 × 10³⁸ possible values, making the probability of collision effectively zero for our use case.

What alternatives have you considered?

• Use the timestamp as the ID: This may fail when multiple queries share the same timestamp.
• Use a group hash code: While useful for identifying groups, it is insufficient for distinguishing individual queries within a group at different timestamps.
• Generate an ID by hashing the query source and timestamp - this won't avoid ID collision neither

Do you have any additional context?

This feature is a prerequisite for opensearch-project/query-insights-dashboards#12

[ansjcy]

We need to pick this up as part of the QID release.

[deshsidd]

As part of this changes need to also incorporate the query_id in the Query Grouping Dashboard PR : opensearch-project/query-insights-dashboards#33

Following PR comments are relevant:

• Query grouping dashboard changes and extensive tests query-insights-dashboards#33 (comment)
• Query grouping dashboard changes and extensive tests query-insights-dashboards#33 (comment)

[deshsidd]

This task comprises of the following subtasks:

1. Model changes to add id
2. Add API for fetching top n for query_id/query_group_id (common id field)
3. Integrate with QI dashboard (display id on the UI)
4. Modify reader to take the id into account when invoking local index for historical data