Improvements

• Account Deactivation Reminders: Search.gov users who have been inactive for more than 90 days must be deactivated, per security requirements. We will now send email reminders to log in prior to the 90 day mark to keep their accounts active.
• Search.gov Account Signup: Users will need to complete all account information in order to manage their websites. On the back end, we have added new fields for first name and last name, which users will soon be seeing in their accounts. Where possible, these fields will be pre-filled, but in some cases, users will need to complete the Last Name field on their next login before proceeding.

Improvements

• Rails Version: We upgraded our search-gov infrastructure to Rails 5.2.
• Infrastructure Updates: We conducted a large portion of housekeeping on our codebase after the Login.gov implementation.

Fixes

• YouTube Results: We addressed a bug that was impacting YouTube search results for some sites using Search.gov.

Improvements

• Rails: We upgraded our application to a new version of Rails
• Security: We upgraded some components of our system to ensure more secure infrastructure.
• Elasticsearch: We continued work on upgrading our analytics engine’s version of Elasticsearch.

Fixes

• Input Improvements: We added input validation to a couple fields that originally allowed off-format submissions.

Improvements

• Rails Upgrade We upgraded our infrastructure to a new version of Rails.
• Elasticsearch Upgrade: We upgraded the Elasticsearch version for our custom search features: Best Bets, Tweets, Videos, RSS, Health Topics, Federal Register, and Related Searches.
• Login.gov integration: Search.gov users logging in will now use Login.gov to authenticate into our platform.
• JQuery Upgrade: We upgraded our JQuery version.

Improvements

• Login.gov: We continued to work on our integration with Login.gov.
• Elasticsearch Indexing: We upgraded the underlying Elasticsearch engine for our optional search features, which includes the following:
  • Best Bets
  • Federal Register
  • Health Topics
  • Related Searches
  • RSS Feeds
  • Twitter
  • YouTube
• Rails Upgrade: We implemented infrastructure changes for our forthcoming Rails upgrade.
• Improved Security: We improved the security and stability of our system through a series of upgrades.

Improvements

• Service upgrades: We upgraded our Rails version and continued working on other internal upgrades that ensured smooth migration across versions.
• Improved security: We resolved and upgraded a series of security vulnerabilities.
• Login.gov: Continued to work on our integration with Login.gov
• Elasticsearch Indexing: We made several changes to prepare for an upcoming Elasticsearch upgrade.

Improvements

• Improved security: We resolved and upgraded a series of security vulnerabilities.
• Service upgrades: We continued to upgrade components of our service, including ElasticSearch, Rails, and Ruby to improve the overall performance of our indexing system.

Fixes

• Account Deactivations: We fixed a bug that erroneously deactivated some users’ accounts due to GSA account use policies.
• Twitter Integration: We fixed a bug that prevented us from retrieving data from the Twitter API.

Improvements

• Increasing indexing coverage We now index pages with only title metadata. Previously, we required pages include either content or description metadata in order to be indexed. This was to ensure we indexed pages with more complete information. However, this limited relevant content included on Javascript pages and similar client-side rendered pages.
  • As you include title metadata, it’s important keep them short and on-topic without packing in repeated information such as agency name. If you want to include your agency name, we recommend you place it after the page title. More information on metadata and tags you should include is available here.
• Improving USAJOBS Integration Common government terms “federal” and “gobierno” no longer trigger our USAJOBS integration. Previously, queries that included these terms triggered our integration with USAJOBS to display relevant postings. However, these queries often resulted in no or irrelevant job postings.
• Service upgrades We continued to upgrade components of our service, including ElasticSearch, Rails, and Ruby to improve the overall performance of our indexing system.

Fixes

• Language parsing bug fix We fixed a bug that parsed non-English pages (e.g. Greek) incorrectly or not at all. This particularly affected pages containing mathematical formulas and equations.

Improvements

• 90 day Account Deactivations: We now deactivate accounts for users who haven’t logged into Search.gov in the last 90 days. If you are unable to login should contact us via search@support.digitalgov.gov to reactivate your accounts. Contractors who do not use .gov or .mil email addresses will be asked to have their government POC re-confirm their status.
• Improved stability: We stabilized search performance by removing unnecessary content from our indices to free up memory.
• Service upgrades: We continued to upgrade components of our service, including ElasticSearch, Rails, and Ruby to improve the overall performance of our indexing system.
• Non-.gov Government Domains Inventory: We decommissioned the old tool we used to manage the inventory of domains in use by federal agencies outside of the .gov and .mil top level domains. We transferred the data to GitHub, where it will be easier for agency staff to contribute to the inventory. More information is available on our website.

Fixes

• Image search result bug fix: We fixed a bug that prevented images sent to our index via Media RSS from being displayed in search results.
• Video search results bug fix: We fixed a bug that prevented videos from appearing in some search results pages.

Improvements

• Sitelimit parameter support
  • Now our customers can specify that a search box only returns results from a particular part of their site. For example, for site www.example.gov, you want a search box on your site to only return results from www.example.gov/history. By adding the appropriate URL parameter and value, you can limit results: &sitelimit=www.example.gov/history.
  • You can also specify multiple values for the sitelimit. See this example from the National Park Service.
  • Keep in mind that the sitelimit must be for a domain within your site’s index. That is, if the Domains list includes www.example.gov, the sitelimit parameter cannot be www.agency.gov/history.
  • This feature is an alternative to our Collections feature.
• Service upgrades
  • We continued to upgrade components of our service, including ElasticSearch, Rails, and Ruby to improve the overall performance of our indexing system.

Fixes

• Highlighting bug
  • We fixed a bug that prevented highlighting for exact, phrased queries on sites indexed by Search.gov. Here’s an example of it working correctly now.