Changelog: added

Changelog: added

Redact named functions parameters when
logging ClickHouse sql queries.

Also update to Bundler 2.6 which is requirement to support checksums
inside lockfiles. Also run `--normalize-platforms`

The version constraint on oj seems very strict, so let's relax it a
little bit, so that the version can be bumped independently in the
main lockfile.

Changelog: added

Changelog: added

Changelog: removed
EE: true

In https://gitlab.com/gitlab-org/gitlab/-/merge_requests/181613 you can
see it was easy to accidentally completely break housekeeper as there is
no integration test for it. That kind of problem (missing a `require`)
cannot be spotted by running RSpec because the order dependence of tests
might (and did) mask the issue.

Adding an integration test here is an easy way to reduce the risk of this
happening again.

The version constraint on json seems very strict, so let's relax it a
little bit, so that the version can be bumped independently in the
main lockfile.

Changelog: added
Signed-off-by: Rémy Coutable <remy@rymai.me>

The version constraint on json seems very strict, so let's relax it a
little bit, so that the version can be bumped independently in the
main lockfile.

Changelog: added

Changelog: added

Model concern allows classes with fixed
items to have active record like methods
(all, where, find_by, find).

The has one association can be used
on models to establish an association to
a fixed items model.

Changelog: added

Changelog: added

In https://gitlab.com/gitlab-org/gitlab/-/merge_requests/181094 we
improved the filtering capabilities in Housekeeper and added a new class
FilterIdentifiers. But unfortunately we did not require the class before
using it and now it's breaking Housekeeper in CI (see
https://ops.gitlab.net/gitlab-com/database-team/gitlab-com-database-testing/-/jobs/17251336
for an example).

You can also easily test this yourself by running any keep. I've tested
locally with the example in the README.md and this fixes the problem.

The version constraint on logger seems very strict, so let's relax it a
little bit, so that the version can be bumped independently in the
main lockfile.

This introduces a comprehensive migration framework for Active Context
and adds collection management capabilities through database-specific executors.
These changes enable systematic database schema management and provide a unified
way to create and manage collections across different database backends.

Key changes:
- Introduce migration framework with version tracking and batch operation support
- Add database-specific executors for managing collections (PostgreSQL, Elasticsearch, OpenSearch)
- Implement partition strategy for handling sharded collections
- Move prefix configuration from Client to Adapter level
- Add migrations path configuration support

Technical details:
- Add Migration base class with versioning support (V1_0)
- Implement BatchOperationResult for tracking long-running operations
- Add collection creation with field type support (bigint, prefix, vector)
- Add partition management with naming strategy
- Add support for creating indices on fields based on database capabilities
- Implement database-specific collection creation logic:
  * PostgreSQL: Partitioned tables with vector support
  * Elasticsearch: Index templates and aliases
  * OpenSearch: Base executor structure (implementation pending)

The migration framework follows a similar pattern to ActiveRecord migrations,
with timestamped migration files and version tracking. Collections can be
created with specific field types and indexing options, which are translated
to the appropriate database-specific implementations.

Migration files should be placed in ee/db/active_context/migrate/ by default,
but this path can be configured through ActiveContext::Config.

We will log response sizes over the GITLAB_LOG_DECOMPRESSED_RESPONSE_BYTESIZE if set

At
https://gitlab.com/gitlab-org/gitlab/-/merge_requests/180120#note_2329196621
it was mentioned that it's difficult to track down exactly where
housekeeper is running. Adding links to the README might help others in
future locating this information.

Changelog: added
EE: true

Changelog: added
EE: true

Changelog: added
EE: true

Changelog: added
EE: true

Implement a comprehensive Elasticsearch query processor that transforms
ActiveContext::Query AST nodes into Elasticsearch DSL queries. This
implementation provides proper handling of all query types and follows
best practices for query composition.

Key changes:
- Add Processor class with support for filter, prefix, knn, and boolean queries
- Implement helper methods for consistent query building
- Add comprehensive test coverage
- Add YARD documentation for all methods

Details:
* Processor changes:
  - Implement transform method following concern contract
  - Add build_bool_query helper for consistent query construction
  - Add proper handling of KNN queries with other conditions
  - Implement proper query composition for nested structures

* Test improvements:
  - Add comprehensive test suite with all query types
  - Add test cases for complex nested queries
  - Add edge case coverage
  - Use shared examples for common query patterns

* Documentation:
  - Add YARD documentation for all public methods
  - Add usage examples
  - Document query composition patterns
  - Add type specifications for parameters

This implementation provides a production-ready solution for transforming
query AST nodes into Elasticsearch queries, with proper handling of all
supported query types and edge cases.

Changelog: added
EE: true

# Conflicts:
#	gems/gitlab-active-context/spec/spec_helper.rb

This reverts commit a1dda13d, reversing
changes made to b10a664e.