Reduce OpenSearch Full reindex with parallel bulk requests

XMLWordPrintable

    • 1
    • CtB - Improve Existing

      Problem

      Currently when performing a Full Reindex, while using OpenSearch with Confluence, the content to be indexed get added to a buffer and then sent to OpenSearch using the _bulk rest api. These bulk requests are currently sent sequentially despite the Full Reindex is performed using multiple threads to process the content.

      Suggested Solution

      Implement parallel bulk requests.

      Why This Is Important

      Parallel bulk requests will decrease the total reindexing time and can use more efficiently the nodes deployed in the OpenSearch data cluster.

      Workaround

      None

            Assignee:
            Unassigned
            Reporter:
            Marco Salvi (Inactive)
            Votes:
            5 Vote for this issue
            Watchers:
            5 Start watching this issue

              Created:
              Updated: