Fix record retry for deleted events on deleted index

Earlier MR: 282 added a retry for Index not found error to delegate index creation only by Indexer service. This works well for create or update event but for delete event, it ends up retrying indefinitely if index is not found. (This could happen in integration tests scenario if index is deleted but delete event is not processed yet).

This change-set removes generic retry on index not found error and adds retry for create or update event only. In case of delete event if index does not exist than it will not retry.