Black Friday is not when you discover your WooCommerce store was fine on average traffic all year. It is when checkout AJAX, cart fragments, and payment webhooks all run at once while marketing sends a fifty-percent-off email. Canadian retailers often stack US holiday promos with local shipping deadlines, so the spike hits even if you are not officially "doing Black Friday."
You do not need a NASA cluster. You need a plan that matches how WordPress actually behaves under load.
What breaks first on WordPress
Usually not the web server static files. PHP and MySQL take the hit:
- Uncached product and cart pages hammer PHP-FPM workers
- Admin and stock updates compete with shoppers for the same pool
- wp-cron and Action Scheduler backlog when CPU is busy
- Payment gateway callbacks time out if checkout is slow
- Search and filter plugins run heavy queries on category pages
A green uptime monitor on the homepage does not mean checkout works. Test the money path.
Four weeks out: baseline and bottlenecks
Run a realistic load test on staging (same PHP version, Redis on/off, same payment gateway sandbox). Note:
- Concurrent checkout sessions before 502s or cart errors
- PHP-FPM
max_childrenwarnings in logs - Slow query log entries on product search
Fix plugin N+1 queries and enable object cache if you have not already. Caching anonymous catalog pages is table stakes; caching cart is not automatic and needs careful rules.
Two weeks out: cache and CDN
Page cache for shop archive and product pages (exclude cart, checkout, my-account). Browser cache headers for assets. CDN in front of images if your origin is a single Canadian VPS in Edmonton or Beauharnois.
Cloudflare orange-cloud is common; verify SSL mode and that WooCommerce sessions still work. Purge rules after price changes so you do not sell last year's SKU at the wrong price.
One week out: freeze risky changes
No major plugin updates, theme edits, or PHP version bumps during peak week unless security-critical. Tag a known-good release in git. Snapshot the VPS or confirm backups restore.
Scale vertically if metrics say you are one tier short: RAM for PHP and MariaDB buffer pool often beats adding complexity.
Day of: operations
- Switch to system cron for WordPress if you still rely on visit-triggered wp-cron
- Disable nonessential admin plugins and marketing popups that add JS weight
- Watch cart error rate and payment gateway dashboard, not just server CPU
- Have a comms plan: status page or social post if checkout degrades
Keep someone who can disable a misbehaving plugin without guessing SFTP paths. That is often the fastest incident fix.
Email and SMS traffic
Your host is not your ESP. Abandoned-cart and promo blasts spike clicks back to the site at the same moment. Stagger sends. Warm up SMTP or use a dedicated transactional provider so order confirmations do not queue behind marketing.
After the spike: learn once
Save graphs from the event: PHP workers, DB connections, queue depth, checkout latency. Those numbers justify next year's hosting budget better than gut feel.
If you limped through on shared hosting, document the hour checkout failed. That is your business case for VPS or managed WooCommerce hosting before the next November.
Bottom line
Black Friday on WordPress is a load test you cannot reschedule. Cache the catalog, protect checkout, test payments under concurrency, freeze changes, and watch the cart, not just the homepage. Canadian stores competing for holiday shipping windows have even less room for a silent checkout failure.
Want a pre-season review of WooCommerce on Canadian hosting? Talk with Swift Host. We will stress-test the boring parts so launch day is less exciting than it has to be.
