feat: Optimize PG source_kind, relation deletion using indices - BED-8832

[StephenHinck]

Description

Implements the new DeleteNodesByKinds/DeleteRelationshipsByKinds functions supported optionally on DAWGS drivers to execute rapid deletion of objects by kind. Initially, this is supported on PostgreSQL drivers.

Requires SpecterOps/DAWGS#101

Note: this was run with the expectation that SpecterOps/DAWGS#100 has already merged. The timings for source_kind deletion include this change.

Motivation and Context

Resolves BED-8832

Why is this change required? What problem does it solve?

How Has This Been Tested?

Validated locally and confirmed existing tests continue to pass.

Screenshots (optional):

Tested using large OGGen payloads for validation:

Delete by source_kind:
Before:

time=2026-06-30T16:15:45.448Z level=INFO message=DeleteCollectedGraphData delete_all_data=false delete_sourceless_data=false delete_source_kinds=OggenBase delete_relationships="" measurement_id=129 elapsed=1m21.810004664s

After:

time=2026-06-30T16:21:37.337Z level=INFO message=DeleteCollectedGraphData delete_all_data=false delete_sourceless_data=false delete_source_kinds=OggenBase delete_relationships="" measurement_id=129 elapsed=1m14.96949745s

Delete by relationship:
Before:

time=2026-06-30T17:17:54.052Z level=INFO message=DeleteCollectedGraphData delete_all_data=false delete_sourceless_data=false delete_source_kinds="" delete_relationships=HasSession measurement_id=129 elapsed=10.140349088s

After:

time=2026-06-30T17:07:53.310Z level=INFO message=DeleteCollectedGraphData delete_all_data=false delete_sourceless_data=false delete_source_kinds="" delete_relationships=HasSession measurement_id=129 elapsed=453.611584ms

Types of changes

• New feature (non-breaking change which adds functionality)

Checklist:

• I have met the contributing prerequisites
  • Assigned myself to this PR
  • Added the appropriate labels
  • Associated an issue: Issues and Pull Requests README #672
  • Read the Contributing guide: https://github.com/SpecterOps/BloodHound/wiki/Contributing
• I have ensured that related documentation is up-to-date
  • Open API docs
  • Code comments (GoDocs / JSDocs)
• I have followed proper test practices
  • Added/updated tests to cover my changes
  • All new and existing tests passed

Summary by CodeRabbit

• Performance
  • Graph data deletion now uses faster bulk operations when supported, which can significantly speed up large deletions.
• Bug Fixes
  • Improved deletion handling for different graph data types, making cleanup more reliable across backends.
  • Preserved existing deletion behavior as a fallback when bulk deletion is not available.

[coderabbitai]

📝 Walkthrough

Walkthrough

DeleteCollectedGraphData is refactored to wrap execution in measure.ContextLogAndMeasure and branch between two deletion strategies: set-based deletes via new nodesByKindDeleter/relationshipsByKindDeleter capability interfaces when the driver supports them, or streaming row-by-row fallback helpers extracted from the prior inline logic.

Graph Data Deletion Refactor

Layer / File(s)	Summary
Orchestration and logging cmd/api/src/daemons/datapipe/delete.go	DeleteCollectedGraphData replaces slog.InfoContext with a deferred measure.ContextLogAndMeasure wrapper and conditionally dispatches to set-based or streaming helpers for both nodes and relationships.
Set-based capability interfaces cmd/api/src/daemons/datapipe/delete.go	Defines nodesByKindDeleter and relationshipsByKindDeleter interfaces; deleteNodesByKindsSetBased maps DeleteAllGraph/DeleteSourcelessGraph/DeleteSourceKinds flags to DeleteNodesByKinds calls while preserving common.MigrationData.
Streaming fallback helpers cmd/api/src/daemons/datapipe/delete.go	deleteNodesByOperation and deleteRelationshipsByOperation encapsulate the prior inline streaming logic: building an operation, reading filtered IDs, and deleting row-by-row via the writer.

Estimated code review effort

🎯 3 (Moderate) | ⏱️ ~20 minutes

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 75.00% which is insufficient. The required threshold is 80.00%.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title clearly matches the PR's main change: optimizing PostgreSQL kind-based graph deletion.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Description check	✅ Passed	The PR description follows the template and includes the required sections, issue reference, testing notes, screenshots, type, and checklist.

✨ Finishing Touches

📝 Generate docstrings

• Create stacked PR
• Commit on current branch

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch BED-8832

---

_{Comment @coderabbitai help to get the list of available commands.}