Elementor 3.17 Update Broke My Forms — Here’s the Fix

Everything looks fine for about twelve seconds. Then the contact forms stop submitting. No error message. No console warning. Just… nothing.

The Real-World Scenario

Happened last Tuesday on a client’s WooCommerce site in Sandton. They’d updated Elementor overnight — standard maintenance window, right? By 9 AM their lead forms were dead. Zero submissions. The spinner just sat there spinning like it was waiting for a bus that wasn’t coming.

What Actually Broke

Elementor 3.17 changed how `action` attributes work on forms. They moved from inline `action=”/wp-json/elementor/forms/submit”` to a dynamic AJAX handler that requires the `elementor_form_submitted` nonce. If your page cache serves a stale nonce? Form dies silently.

Three things I check in order:

1. Clear all caches — WP Rocket, LiteSpeed, Cloudflare, server-level (HostAfrica/afrihost/nginx). The nonce in the cached HTML is from before the update.

2. Regenerate Elementor data — Elementor > Tools > Regenerate Files & Data. This rebuilds the form endpoints with fresh nonces.

3. Check for `form_data` in `localStorage` — 3.17 caches form state locally. If the user had the page open during update, their browser sends stale data. Hard refresh (Cmd+Shift+R) fixes it.

The Johannesburg Angle

We see this constantly on sites hosted with Afrihost, Web Africa, and Hostigator. Why? Their server-level caching (nginx fastcgi, LiteSpeed LSCache) serves stale HTML for hours after an Elementor update unless you purge manually. Most clients don’t know their host caches at the server level — they think “I cleared WP Rocket, I’m good.”

Had a client on Web Africa Business Hosting last month. Cleared WP Rocket, cleared Cloudflare. Forms still broken. Turns out Web Africa’s nginx cache has a 4-hour TTL by default. Had to log into their cPanel > LiteSpeed Cache > Purge All. Forms worked instantly.

Pro Tip from the Trenches

Disable Elementor’s “Improved Asset Loading” if you’re using ANY caching plugin. They fight. Elementor tries to optimize CSS/JS delivery while WP Rocket/LiteSpeed does the same thing differently. Result: broken nonces, missing CSS, forms that submit to the wrong endpoint.

We’ve seen this cause 3+ second delays on Johannesburg-hosted sites specifically — the asset loading rewrite adds overhead that server-level caching then serves to everyone.

Final Thoughts

Elementor 3.17 isn’t “broken” — it changed the contract. Your cache is serving the old contract. Purge everything, regenerate Elementor data, hard refresh. Takes 3 minutes. Your leads will thank you.

One more thing. The stale-cache half of this has a reusable answer. The nonce half is an Element Caching setting, and I wrote up the real cause properly here.

Read the corrected version