perf: add database indexes on frequently queried columns

[Fayyo]

Summary

Add database indexes on frequently queried columns to improve query performance as the dataset grows. Indexes target the most common query patterns identified in route handlers: listing expenses by group (sorted by date), fetching settlements ordered by date, and retrieving audit logs by entity.

Changes

New indexes added to prisma/schema.prisma

Model	Index	Columns	Purpose
Expense	@@index([groupId, createdAt])	group_id, created_at	Composite index covering the primary list query findMany({ where: { groupId }, orderBy: { createdAt: "desc" } }) used in expense listing, ledger, and balance computation
Settlement	@@index([createdAt])	created_at	Supports ordering by date in history (GET /history) and worker FIFO processing
AuditLog	@@index([entityType, entityId])	entity_type, entity_id	Composite index for filtering audit logs by entity type and ID
AuditLog	@@index([createdAt])	created_at	Supports ordering audit logs by timestamp

Existing indexes that already covered other AC requirements

• GroupMember: @@index([userId]) and @@unique([groupId, userId]) (composite B-tree covers groupId, userId lookups) ✅
• Expense: @@index([groupId]), @@index([payerUserId]) ✅
• Settlement: @@index([groupId]), @@index([status]), @@index([fromUserId]), @@index([toUserId]) ✅
• AuditLog: @@index([userId]) ✅

Migration

Generated migration at prisma/migrations/20260728000001_add_query_indexes/migration.sql:

Index	SQL
Expense group+date	CREATE INDEX "expenses_group_id_created_at_idx" ON "expenses"("group_id", "created_at");
Settlement date	CREATE INDEX "settlements_created_at_idx" ON "settlements"("created_at");
AuditLog entity	CREATE INDEX "audit_logs_entity_type_entity_id_idx" ON "audit_logs"("entity_type", "entity_id");
AuditLog date	CREATE INDEX "audit_logs_created_at_idx" ON "audit_logs"("created_at");

Note on over-indexing

Expense does not have a status field (status exists on ExpenseShare, not Expense itself), so the suggested @@index([status]) was omitted to avoid referencing a non-existent column.

Testing

• ✅ All 79 tests across 11 test files pass with no regression
• ✅ Indexes verified present in database via \di
• ✅ Migration is non-destructive and safe for production (no downtime)

Closes #35

[drips-wave]

@Fayyo 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