Druid 27.0 release planning

[2bethere]

This issue should cover things we want to land for the upcoming Druid 27 release and to coordinate any dependencies/reviews/docs and testing.

Milestone tracking

ETA for code free is July 7th, 2023
https://github.com/apache/druid/milestone/51

Basing this on

• Community roadmap 2023 #14157

Features

• Concurrent compaction + ingestion (--> Moving to Druid 28)
• Window functions (Experimental)
• Cold-tier/Async query (Experimental)
• Simplify segment balancing for improved stability
• Schemaless ingestion (GA)

Dependency changes

• Hadoop 3 as the default build ([Backport] docs: add docs for schema auto-discovery (#14065) #14301 )
• Hadoop 2 deprecation (Announcement, with depreciation targeting Druid 28)
• Update to Calcite 1.34 or later (Upgrade calcite in druid #13532 --> Moving to Druid 28)

Bug fixes and paying down tech debts

• Adding SQL null by default mode wrapper for future releases
  • SQL-compatible null handling and strict booleans by default #14154
  • Properly read SQL-compatible segments in default-value mode. #14142)
• Official support for jdk 17 (Claim full support for Java 17. #14384)
• bug fixes

[2bethere]

@techdocsmith FYI