WooCommerce database load: find the queries and tables slowing a store

Separate storefront, checkout, admin and background load

Measure request latency/query time for product pages, search/filter, cart/checkout, account, wp-admin orders/reports, and Action Scheduler jobs. Different paths stress different tables and cache layers.

A slow admin report should not be diagnosed from storefront cache-hit performance.

Capture repeated and slow queries

Use database slow-query tooling and WordPress profiling to identify execution count, rows examined, indexes, locks, and N+1 patterns. Extensions may add joins against postmeta/order metadata that dominate specific pages.

Compare query plans after WooCommerce storage-model or plugin changes.

Use cache only where correctness permits

Page caches should bypass cart/checkout/account and personalized responses. Object cache can reduce repeated reads but does not remove expensive uncached write paths or bad queries.

Invalidate product/stock/pricing data correctly when using persistent cache.

Plan capacity around peaks and writes

Sales/imports can create concentrated order and inventory writes. Monitor DB CPU, memory, disk latency, connections, lock waits, and PHP worker queues during peak scenarios before resizing. Capture a normal weekday baseline as well as a promotion/import peak so a short burst is not mistaken for permanent capacity demand.

• Profile distinct WooCommerce paths.
• Find slow/repeated queries.
• Bypass personalized page cache.
• Measure write-heavy peak behavior.