HDDS-12643. Fix ContainerTable pagination in Recon UI v2

[arunsarin85]

What changes were proposed in this pull request?

The Recon UI v2 ContainerTable had three pagination bugs reported in HDDS-12643:

1. All data was fetched in a single request and paginated client-side; lastKey was never used as a cursor.
2. Users could jump directly to any page number.
3. The row limit was not configurable.

Please describe your PR in detail:

This PR fixes all three issues:

1. Server-side pagination (containers.tsx): Replaced the single bulk-fetch + client-side slice approach with true server-side pagination. Each page fetch sends limit (page size + 1) and minContainerId to /api/v1/containers/unhealthy/{STATE}. Fetching one extra item is used solely to detect whether a next page exists (hasNextPage = response.length > pageSize), and only pageSize rows are ever displayed. Each tab maintains independent pagination state so switching tabs does not reset other tabs.
2. No direct page-skip (containersTable.tsx): Replaced Ant Design's built-in numbered pagination with a custom footer containing only Previous and Next buttons. Direct page number access is no longer possible.
3. Configurable page size (containersTable.tsx, containers.tsx): Added a Rows per page dropdown (options: 10, 25, 50, 100) in the footer. Changing the page size resets all tab states and re-fetches the first page with the new limit, and is reflected across all tabs.

What is the link to the Apache JIRA

https://issues.apache.org/jira/browse/HDDS-12643

How was this patch tested?

1. Mock API (manual): Ran npm run dev with the updated mock:api script (which now includes --middlewares api/pagination.js). Verified via browser DevTools that:

• Each page click issues a new GET request with the correct minContainerId cursor and returns a 200 (not 304) with a different set of container IDs.
• Next is disabled on the last page; Previous is disabled on the first page.
• Changing the Rows per page dropdown resets to page 1 and issues a new fetch with the updated limit.
• All 5 tabs (Missing, Under-Replicated, Over-Replicated, Mis-Replicated, Mismatched Replicas) paginate independently.

2. Random-ID validation: Container IDs in db.json are stored in shuffled (non-sorted) order with non-sequential values. All 5 tabs returned correctly sorted, correctly paginated results, confirming the middleware's Array.sort step handles arbitrary ordering in the source data.

Did some additional validation

1. Created the missing and under-replicated containers and validated the below scenarios :
   1.a. The Highlights tab values are populated from the initial missing_container API call .
   1.b By clicking on the other tabs - Under_Replicated, Over-Replicated, Mis-replicated the values in the highlight tab is not getting cleared out

[spacemonkd]

Thanks for the continuous effort on this patch @arunsarin85.
This looks good to me now, +2

[adoroszlai]

Thanks @arunsarin85 for the patch, @spacemonkd for the review.