In the end, implementing a custom indexing queue turned out to be too complex. Instead, improvements were made based on the existing queue: * The number of remaining documents is displayed * The indexed chunks are displayed in a Live Data * When an error occurs during embedding, the error is displayed in the list of indexed chunks * Embedding models allows specifying the supported parallelism now and this is taken into account by the indexing process. However, only chunks of a single document are processed in a single request.
Re-indexing of documents and collections still isn't supported and will be implemented as part of a new Jira issue LLMAI-92.
This message was sent by Atlassian Jira (v9.3.0#930000-sha1:287aeb6)
If image attachments aren't displayed, see this article.
