Performance Optimization

Optimize for reads. Do work at write time, not read time.

Counter Caches

Use when you frequently count associated records.

# Migration
add_column :prompts, :comments_count, :integer, default: 0, null: false

# Model
belongs_to :prompt, counter_cache: true

Denormalization

Pre-compute values and store them directly. Update via background jobs when source data changes.

# Instead of computing on every read:
def total_revenue
  orders.sum(:amount)
end

# Store it:
add_column :shops, :total_revenue_cents, :bigint, default: 0

Decision Framework

Before Optimizing

Always check simpler solutions first:

1. •Can includes fix it? - N+1 queries are the most common issue
2. •Is there an index? - Missing indexes cause full table scans
3. •Is the query optimal? - Check EXPLAIN ANALYZE
4. •Is it actually slow? - Profile before optimizing

Caching Patterns

# Fragment caching
<% cache shop do %>
  <%= render shop %>
<% end %>

# Low-level caching
Rails.cache.fetch("shop/#{shop.id}/stats", expires_in: 1.hour) do
  compute_expensive_stats
end

Common Anti-Patterns

• •Premature optimization without profiling
• •Caching without invalidation strategy
• •N+1 queries in loops
• •Loading full objects when you only need IDs