[META] [Star Tree] [Search] Supported Queries to resolve via star tree

[sandeshkr419]

Please describe the end goal of this project

This issue keeps a tracker of supported aggregations & query shapes which can be resolved via star-tree composite indices and the star-tree implementation (if possible) should benefit the performance for the query shape.
This will basically keep a track of supported and planned query shapes which can utilize star-tree aggregation for faster computation.

As for 2.17 release, only numeric fields are supported in star-tree composite indices. The list will be updated as we identify more query shapes that we can pick up.
Community - Please do share/comment on what all you think should be prioritized.

Here is the brief segregation on milestones:

Milestone 1: The primary goal of Milestone 1 is to utilize star-tree to resolve metric aggregation queries (max, min, avg, sum, value count) over numeric fields along with support of a single query as well. This milestone lays out the building blocks of how star-tree approach can be integrated in within a search request to resolve it.

Milestone 2: The primary goal of Milestone 2 is support certain queries within above aggregation requests like term/terms query, boolean query over numeric/keyword fields.

Milestone 3: The primary goal of Milestone 3 is to support single-level bucket aggregations like date histogram, range aggregations, single bucket aggregations, etc

Milestone 4: The goal of this milestone is to extend the star-tree support to resolve nested aggregations, multi-terms aggregations.

Milestone 1 (2.18)

• Metric Aggregation
  [Star Tree][Search][RFC] Parse aggregation request to resolve via star tree data structure #14871
  [Star Tree] [Search] Support for metric aggregations with/without term query #15289
• Metric Aggregation with numeric field term query
  [Star Tree][Search][RFC] Parse aggregation request to resolve via star tree data structure #14871
  [Star Tree] [Search] Support for metric aggregations with/without term query #15289

Milestone 2 (2.19)

• Metric Aggregation with keyword field term query
  [Star Tree] [Search] Support for keyword field in term query with metric aggregations #16537
• Metric Aggregation with IP field term query
  [Star Tree] [Search] Support for IP field in term query with metric aggregations #16547
• Metric Aggregation with multi-terms numeric/keyword query
  [Star Tree] [Search] Support for terms query with metric aggregations #16538
• Metric Aggregation with numeric range query
  [Star Tree] [Search] Support for numeric range query with metric aggregations #16539

Milestone 3 (2.19)

• Date histogram with metric aggregation
  [Star Tree] [Search] Date histogram with metric aggregation #16552
  [Star Tree] [Search] Resolving Date histogram with metric aggregation using star-tree #16674
• Bucket terms Aggregation with metric aggregation
  [Star Tree] [Search] Bucket terms Aggregation with metric aggregation #16551

Milestone 4 (TBD)

• Range aggregation with metric aggregation
  [Star Tree] [Search] Range Aggregations with metric aggregation #16553

• Multi-terms aggregation

• Nested bucket aggregation

Related component

Search:Aggregations

[expani]

@sandeshkr419 @getsaurabh02 Kindly assign this to me.

[sandeshkr419]

Capturing post 2.19 items to investigate:

• Optimize numeric conversions - [Star Tree] [Search] Keyword & Numeric Terms Aggregation #17165 (comment)
• Utilize star-tree in bucket aggregations without any sub-aggregations - [Star Tree] [Search] Resolving Date histogram with metric aggregation using star-tree #16674 (comment)
• Evaluate correctness of DEPTH_FIRST execution in bucket aggregations - [Star Tree] [Search] Keyword & Numeric Terms Aggregation #17165
• Evaluate if deferrable collectors also need to extend pre-computation framework