pn-audit-performance

pn-audit-performance

Start every response with: [pn-command] 🔺

Advanced (palette-hidden). Not in the / slash palette. Invoked by the /pn-backend-audit umbrella, or directly via get_command("pn-audit-performance").

Focused backend performance pass: identify and fix N+1 queries, missing caching, unoptimized DB queries, blocking I/O, and resource exhaustion. No API design changes, no security changes — performance only.

Flow

1. Context

Check .pncore-stack.md for DB, ORM, caching layer, and hosting environment. If not found, ask:

• "What database and ORM? (PostgreSQL + Prisma / MySQL + Sequelize / etc.)"
• "Any caching in use? (Redis / Memcached / in-memory / none)"
• "Any observability showing slow queries? (Datadog / Sentry / logs / no visibility)"

2. Scope

If not specified: "Which routes, services, or data access patterns should I focus on? Or reply 'all' for a full pass."

3. Audit

Load get_skill("pn-caching") and consult pn-core://skills/backend/reference/database-patterns.md.

N+1 queries:

• Loops calling DB inside each iteration?
• ORM lazy loading associations accessed in a response handler?
• Repeated identical queries with different IDs visible in logs?

// N+1 pattern to look for:
for (const order of orders) {
  const user = await db.users.findById(order.userId); // N+1
}

// Fix: batch or JOIN
const orders = await db.orders.findAll({ include: ['user'] }); // ORM eager load
// OR: SELECT o.*, u.* FROM orders o JOIN users u ON u.id = o.user_id

Missing indexes:

• FK columns without indexes?
• High-traffic WHERE columns without indexes?
• ORDER BY with LIMIT without index on sort column?
• Queries using sequential scans on large tables? (EXPLAIN ANALYZE results)

Caching opportunities:

• Frequently-read data that rarely changes? (config, user permissions, product catalog)
• Expensive aggregations recalculated on every request?
• External API calls made on every request for slowly-changing data?
• Session data loaded from DB on every authenticated request?