Release Notes 4.14.x

Release 4.14.0

New Feature

  • Search deduplication:  when this is enabled, if you search for an object that exists in multiple places (e.g. a table duplicated across different data sources and schemas), you’ll see the different matches grouped together and have the ability to pick the particular instance you want. If you’re interested, talk to your admins and Solutions Engineer.

Improvements

  • Now, Table Pages for MySQL and MemSQL Data Sources will show Update Time.

Bug Fixes

  • Values on Column pages are displayed in order of descending  frequency, so the most frequent values in the column are shown up top. This worked historically but then regressed.

  • Previously, when scheduled queries with comments would run, they’d sometimes be reported as having failed even when execution succeeded. This will no longer happen.

  • Scheduled queries will now run with the specified custom connection-string, rather than the data source default.

  • In Compose, the Explain menu now works for Netezza.

  • Throughout Catalog (in both built-in and custom fields) users who’ve been suspended by an admin will be indicated by ‘[GONE]’ being affixed to their names.

  • Users’ job titles now appear next to their names in Tagged Objects lists.

  • Clicking the login button twice in a row will no longer cause authentication issues.

  • @-mentions now work with non-English keyboards.

  • Flags on published queries will now display properly.

  • Alation will no longer add arcs to Lineage  for failed CREATE statements,

    • and if you successfully run multiple CREATE statements in Compose during a single query execution, arcs will be added for each. Previously, only the last would be considered.

  • It is now possible to read text in the Tags field on the Description pane in Compose when using the Dusk Theme. Previously the text was white-on-white.

  • The in-app user satisfaction survey

    • now displays the full question without truncation

    • and accepts non-ASCII characters in the feedback form.

  • The Ask-a-Colleague modal now works better. Previously, certain flows would cause two overlapping dialog boxes to appear simultaneously.

  • Query counts on Tables pages are now computed and displayed correctly.

  • SmartSuggest will now work better for all supported data source types. Previously, keyword lists for certain sources/dialects, such as MemSQL, were unavailable and caused problems.

  • Images and text will no longer bleed across boundaries in small windows.

  • Tags are now once again searchable by their prefixes. A regression in 4.10.x made it possible to find tags only by exact name match.

  • We now show a peek when hovering over the column type in the sidebar on the right of a Column page in Catalog. This allows you to see the full value for the column type definition that can be long in the case of complex nested types. However, we limit the capture of this string to 100 characters, so longer type definitions will still be truncated.

Admin Release Notes

New Features

  • Automatic Extraction for Amazon S3

    • Got a Data Lake in Amazon S3 you want to curate? Alation can now automatically crawl and catalog directories and files in S3 Buckets, creating pages for each asset where you can flag, tag, title, describe, and otherwise annotate them.

  • Archiving Old Ingested Queries  to save space

    • Now you can configure query logs that have already been processed to be archived or deleted. This can save lots of disk space on your instance. You can find this option in Admin Settings > Misc at the bottom in the section titled “Archive processed QLI data”.

    • Note: Query logs will only be archived if they are older than the given time interval. Note that the logs are archived or deleted in chunks of about 10GB so you will not be able to archive anything if you have less than this threshold worth of query log data.

  • And we now connect to Spark SQL and SAS DB. You can extract metadata for and execute queries against each.

Improvements

  • It is now possible for Stewards to create new Custom Templates with no fields that can be helpful for classifying articles and making them facet-searchable.

  • We’ve also improved our API documentation for GET Metadata (including from Custom Fields and Report Sources) and for Lineage.

Bug Fixes

  • It is now possible to add a BI Source with a URI that Alation has seen before. Previously this was not possible, even when the old source had been deleted.

  • The GET Metadata API for rich-text fields will now return full values, including HTML tags—so the API result matches what users see in the UI. Previously, the API returned values truncated to 20 characters with HTML tags stripped away.

  • Data Profiling now works for tables containing column names that are also reserved keywords in the data source’s SQL dialect.

  • Previously, instances with SAML authentication ignored the “Moderate user signups” checkbox in settings. Now if that box is checked, new users will need to be activated by an admin before they can access Alation. Users who had already signed into Alation once will continue to be able to access Alation without needing to be activated.

  • Metadata Extraction robustness:

    • SQL Server: extraction now works when a particular database is specified in the interface and when certain parameters (e.g. UseNTLMv2) are included in the URI

    • Snowflake  extraction will no longer fail due to a serialization error

    • Teradata Aster columns on VIEWs will be extracted. Previously, TABLE columns were included but VIEW columns were missed.

  • We’ve fixed a regression introduced in 4.12 that broke Greenplum Query Log Ingestion so it couldn’t run at all. Now you can again.

  • Previously a bug made it impossible to update Object Set-, People Set-, or Reference-type fields in Customize Catalog if a group (rather than an individual user) had been added as a value to that Custom Field on any object. That’s been fixed.

  • Allie (the Alation AI bot) can now populate the Lexicon (mappings from literal tokens in data-object names like ‘rvnu’ to human-readable terms like “revenue”) even when some entries have non-ASCII characters. Previously she’d choke on Unicode characters.

  • We now clean up temp files created during Query Log Ingestion even if QLI crashes before finishing.

  • Our ER Studio integration now supports stripping out Rich Text Formatting in descriptions. This feature can be turned off by running the following command from the Alation backend shell: alation_conf alation.feature_flags.strip_rtf_characters_in_erstudio_descriptions -s False

  • Deleted MicroStrategy sources will now be labeled as such. Previously, they looked identical to normal sources, which made it appear that deletion didn’t even work.

    • Also, deleted BI sources (for any Source type) can now be restored.

  • Over 100 data sources can now be viewed in the Deployment Command Center. Previously, only 100 would be shown.

  • It is now possible to profile Teradata Aster  tables and columns with names containing space (’ ‘) or period (‘.’).

  • It is now once again possible for System Administrators of Alation to specify a post-process script for backups (e.g. to move them to a remote server).

Release 4.14.1

Bug Fixes for End-Users

  • Icons on page headers will now display correctly in Firefox. Previously, they’d appear cropped in that browser.

  • When users mouse over a button that’s disabled, they’ll see a tooltip explaining why (often this has to do with their individual permissions). This used to work but regressed.

  • In Compose, the run log for an in-progress Hive query will not disappear even if the next Hive request times out. Previously a time-out would wipe away earlier lines of the log.

  • Hive log requests no longer queue up and consume Alation connector resources if your Hive data source is slow to respond with the log. This should eliminate crashing that sometimes occurred when the works were gummed up.

Bug Fix for Admins

  • LDAP Group syncing now works for non-ASCII characters. Previously, Unicode characters caused failures.

    • Also,  we fixed a regression introduced in release 4.11 that caused LDAP Groups to not sync properly.

Release 4.14.2

Bug Fixes for End-Users

  • Reply-by-Email now works from Outlook 365. Previously, messages sent from Outlook 365 would generate blank posts in Alation.

  • Column metadata from Views are now extracted and shown for SAP HANA datasources. Previously, only column metadata from Tables was extracted from SAP HANA datasources.

Release 4.14.3

Bug Fixes for Admins

  • Alation can now extract MicroStrategy BI Servers that have non-English names for standard folders Public and Reports, if MicroStrategy Admins specify a Windows shortcut. Previously, Alation could not extract from such servers.

  • Pending Users: customers with very specific situations may need to upgrade to 4.14.3 promptly. Your Customer Success Manager has reached out to your Admin if your company may be affected.

Release 4.14.4

Admin Release Notes

New Feature

  • Compose Support for Hive/Presto-like Virtual  Data Sources

    • it is now possible to configure a Virtual Data Source and Connection String combination to let users write and run queries in Compose against systems like Qubole, which Alation does not yet automatically extract.

Bug Fixes

  • Hive Query Log Ingestion will now continue if some summary files are missing. Missing files will simply be skipped. Previously a missing summary file would cause the entire ingestion job to fail.

  • Data profiling in Hive data sources will work when table and column names are also reserved keywords. This support was included in 4.14.0 for other data sources.

Release 4.14.5

Bug Fixes for Admins

  • Alation can now extract MicroStrategy BI Servers that have non-English names for the standard folders Public Objects and Reports, if Alation Sys-Admins run a backend command as described below. Previously (in 4.14.3), Alation could only extract from such servers only if MicroStrategy Admins added a Windows shortcut. To set up, run:

    • alation_conf alation.microstrategy.default public folder_name -s <yourpublicfolder>

    • alation_conf alation.microstrategy.default report folder_name -s <yourreportfolder>

  • Metadata Extraction for Snowflake is working once again. A change to the Snowflake JDBC driver broke extraction in 4.14.0.

  • Query Log Ingestion for Presto 0.182 now works, due to improvements to Alation’s Parser.

Release 4.14.6

Bug Fix for End-Users

  • It is now possible to add a Flag (Endorsement, Deprecation, or Warning) to an object when Custom Permissions have been set for Flags, as long as you’re in the permitted group. A regression in 4.12.0 made it so custom permissions prevented anyone from adding Flags.

Bug Fixes for Admins

  • You can now once again go to the Authentication tab in Admin Settings. A regression in 4.14.0 made caused a 500 error when attempting to visit that page.

  • We’ve added a backend configuration for excluding schemas with a large number of tables from SmartSuggest. For data sources with very large schemas (in this case, over 100k tables), we can configure a limit so that those schemas are excluded that can alleviate crashes on the server or memory bloat on the client.

Release 4.14.7

Bug Fix for End-Users

  • Tableau lineage and SQL Server fix - Resolved issue where users were not able to see lineage all the way to the original table.

Bug Fixes for Admins

  • Unable to configure MSSQL datasource when using Active Directory credentials - Patch to MSSQL driver applied to support AD credentials.

Release 4.14.8

Improvements for Admins

  • Alation now supports HTTPS proxy for sending usage and health data to Alation.

Release 4.14.9

Bug Fixes for Admins

  • Improved performance and hardened MDE extraction for Oracle.

Release 4.14.10

Bug Fixes for Admins

  • Fixed bug where Greenplum and Postgres extraction skipped tables that had columns with comments (in the source database) but the table did not.