7.12.0 release notesedit

Upgrading to Enterprise Search 7.12.0? See Upgrading from Enterprise Search 7.11 and earlier.

Enterprise Searchedit

New featuresedit

  • Optimizes document storage:

    • Up to 40% improvement in document ingestion times.
    • Up to 80% improvement in storage requirements.
  • Removes the need for re-indexing after schema changes (schema changes are now instant).
  • Improves query precision:

    • Increased query precision with fewer long-tail results.
    • Improved typo-tolerance for more resilient search.
    • This change may increase the CPU load on Elasticsearch for some deployments. Beginning in Enterprise Search 7.13.0, you can disable advanced typo-tolerance (fuzzy matching) to reduce CPU usage. To do so, set the query precision value to 8 or higher. See Precision tuning in the App Search documentation.

Breaking changesedit

  • Additional storage capacity is temporarily required while upgrading. See Upgrading from Enterprise Search 7.11 and earlier. Once completed, certain indices can be purged, resulting in a smaller, more efficient deployment.
  • Query precision improvements and improved typo-tolerance mean you may see different search results compared to previous versions.

Bug fixesedit

  • Fixed an issue with session cookies that could cause users in some browsers to be logged out prematurely.
  • While communicating with Elasticsearch, Enterprise Search is now more resilient to network errors that did not originate from Elasticsearch.

App Searchedit

New featuresedit

  • Crawls are performed in a multi-threaded fashion now (10 threads per crawl by default, configured via crawler.crawl.threads.limit), which significantly increases performance of a crawl (up to 5-10x in a default configuration).
  • Adds robots.txt support to the web crawler. See Robots.txt file in the web crawler reference.
  • Adds sitemaps support to the web crawler. See Sitemap in the web crawler reference.
  • Adds more fields to the web crawler document schema, on which you can search, sort, and filter documents. See Web crawler schema in the web crawler reference.

    • url_scheme
    • url_host
    • url_port
    • url_path
    • url_path_dir1
    • url_path_dir2
    • url_path_dir3
    • headings
  • Adds more configuration settings for the web crawler. See Web crawler configuration settings in the web crawler reference.

    • Request timeouts and limits:

      • crawler.http.connection_timeout
      • crawler.http.read_timeout
      • crawler.http.request_timeout
      • crawler.http.redirects.limit
      • crawler.http.response_size.limit
    • Content extraction resource limits:

      • crawler.extraction.extracted_links_count.limit
      • crawler.extraction.title_size.limit
      • crawler.extraction.body_size.limit
      • crawler.extraction.keywords_size.limit
      • crawler.extraction.description_size.limit
      • crawler.extraction.indexed_links_count.limit
      • crawler.extraction.headings_count.limit
    • Crawl parallelization limits:

      • crawler.crawl.threads.limit

Breaking changesedit

Bug fixesedit

  • Allows canceling a crawl during the purge/deletion phase.

Workplace Searchedit

New featuresedit

  • Improved memory usage while indexing. Now makes smarter use of threads on systems with more memory which increases indexing speed. Also fixes some issues that were causing Workplace Search to run out of memory while indexing.

Breaking changesedit

  • Deprecates configuration setting workplace_search.custom_api_source.document_size.limit.

    Use workplace_search.content_source.document_size.limit.

  • Deprecates configuration setting workplace_search.custom_api_source.total_fields.limit.

    Use workplace_search.content_source.total_fields.limit.

Bug fixesedit

  • Deleting a content source now quickly removes any synchronization jobs for that content source.
  • Fixed a race condition that could cause the search experience to break when using a Slack content source.
  • Surfaced an error when a Custom API source field was given a name that was already in use.

Known issuesedit

  • Custom Sources that specify the visible fields that appear in result details are broken by the upgrade to 7.12.0. The issue will manifest as being unable to load the Sources page and search results from Custom Sources appearing malformed. It is recommended to wait until 7.12.1 before upgrading if you use Custom Sources with configured visible fields. This issue does not affect fresh installations of 7.12.0.
  • The Slack connector will not work correctly for OAuth apps created after February 24, 2021. This is caused by the breaking changes that were introduced in the Slack API.