Skip to main content

Imply Polaris release notes

These release notes provide information about the latest features, improvements, and fixes for Imply Polaris.
For older release notes, see the archive.
For information about Imply Enterprise, see Imply Enterprise and Hybrid release notes.

Polaris version

The release notes for Polaris are organized by release date. It may take a few days from the release date for the new Polaris version to become available for your projects.

You can verify the current Polaris version identifier for your projects in the Polaris UI or the Projects API.

May 20, 2026​

May update | Polaris version 50370fe9

Improvements​

  • You can now pass an optional ISO 8601 interval to the PIVOT_QUERY_INTERVAL function as a fallback default time filter. This prevents the function failing if Polaris can’t determine the time filter from the query (id: 73439)
  • Polaris now clips timeseries aggregators to the segment interval. This is particularly useful for buffer and vector aggregators, which need to allocate resources for all possible time buckets (id: 73793)
  • Extended segment-optimized aggregators to groupBy queries. To disable this optimization for debugging purposes, set the optimizeAggregators query context parameter to false (id: 73616)
  • Added Asia/Dubai to the timezone drop-down in data cube general settings (id: 73327)

Fixes​

  • Fixed an issue with reports configured to deliver on the Nth weekday of the month—for example, first Monday of the month. Polaris now delivers these reports on the correct schedule (id: 73642)
  • Fixed an issue that caused asynchronous query results to prematurely disappear (id: 71361)
  • Fixed an issue with multi-topic Kafka ingestion (id: 73225)

April 9, 2026​

April update | Polaris version 8cc65208

Fixes​

  • Fixed an issue with queries that used expression filters or expression-based aggregators when there was a multi-value dimension (id: 72776)
  • Fixed an issue where Kinesis task live reports didn't work (id: 72777)

March 16, 2026​

March release | Polaris version 34caec1d

Fixes​

  • Fixed an issue where compaction failed if an ordering dimension had no corresponding column. Compaction tasks now ignore such dimensions (id: 72228)
  • Fixed an issue with parsing nested columns containing empty fields (id: 72196)
  • Fixed an issue where the Router incorrectly reported successful MSQ task requests as failed. The Router now treats all 2XX response codes as successful, not just 200 (id: 72195)
  • Fixed an issue where realtime ingestion tasks could fail with "inconsistent metadata" errors during auto-scaling (id: 71879)

February 25, 2026​

February release | Polaris version 0b903113

Improvements​

This release includes all performance enhancements and bug fixes from Apache Druid® 36.0.0. For details, see the Druid 36.0.0 release notes.

January 28, 2026​

January release | Polaris version 58c1a667

Band chart visualization​

The new band chart visualization allows you to display a primary metric as a line within upper and lower bound measures. You can use it to identify when values fall outside expected ranges or to show the relationship between a central measure and its constraints (id: 69939)

Improvements​

  • The SQL workbench now supports the Dart engine as a query execution option when Dart is enabled in Druid (id: 70566)
  • Added custom measure PIVOT_QUERY_INTERVAL. It returns the ISO 8601 interval string that covers the evaluated time filter for a query (id: 71221)
  • Improved performance of the DOWNSAMPLED_SUM_TIMESERIES aggregation function (id: 69170, 71349)

Fixes​

  • Fixed an issue with filtering jobs by connection name (id: 70676)