feat: aggregate user retention cohorts in set queries

[shogun444]

Description

Refactors KpiService.calculateUserRetention() to replace nested per-cohort queries and unbounded IN (:...cohortUserIds) clauses with set-based grouped queries, preventing database strain and scaling bottlenecks.

Closes #1535

Changes

• Set-Based Aggregation (src/utils/masking/kpi.service.ts):
  • Replaced userRepository.find() and per-cohort nested loops with 2 grouped aggregate queries:
    1. Grouped cohort counts across precomputed cohort windows.
    2. Inner-joined user retention events grouped by (cohort_month, retained_month).
  • Maintained exact date boundaries using JS parameters to prevent calendar vs rolling 30-day discrepancies.
  • Preserved existing Prometheus gauge labels (cohort_month, retained_month) and values.
• Unit Tests (src/utils/masking/kpi.service.spec.ts):
  • Updated mocks to cover createQueryBuilder aggregate chains (innerJoin, groupBy, getRawMany).
  • Verified retention rates and confirmed the removal of IN (:...cohortUserIds) calls.

Acceptance Criteria Checklist

• Retention computed without per-cohort nested query loops or unbounded IN-lists
• Reported retention values remain identical
• Unit tests updated and CI passing
• Starred the repository

[drips-wave]

@shogun444 Great news! 🎉 Based on an automated assessment of this PR, the linked Wave issue(s) no longer count against your application limits.

You can now already apply to more issues while waiting for a review of this PR. Keep up the great work! 🚀

Learn more about application limits

[RUKAYAT-CODER]

Thank you for contributing to the project.