-
Notifications
You must be signed in to change notification settings - Fork 0
Update and Save Analysis
Profiles the administrator's real Update or Save as the write request itself.
The save you just pressed, measured: request phases, per-plugin cost, SQL, outbound HTTP, mail, peak RAM and, with excimer installed, the functions by self time. The same diagnostics as any other profile, on a request that normally leaves no trace.
It works on existing post, page, product, custom post type and WooCommerce HPOS order edit screens.
WordPress saves over ajax or REST and then reloads the editor. A tool watching the page render sees the editor screen loading afterwards, not the write. So a save_post callback that takes twenty seconds hides in a request nothing was watching, and the symptom the shop owner reports is "saving a product is slow" with nothing to point at.
This profiles the write request itself, and the editor page that loads afterwards is explicitly excluded. A slow save_post callback can't hide inside an ordinary edit-screen page profile.
Not a temporary duplicate, not a dry run. Your edits are saved. Normal mail, webhooks and integrations run unchanged and are measured, which is the point: a save is slow because of what it sets off.
Two transports are handled. The classic editor's authenticated post.php form POST, and the block editor's authenticated REST update. Classic form tokens are removed from both $_POST and $_REQUEST before ordinary plugins load, and REST tokens travel in the signed profiling header, so no plugin sees a stray token in the request it's handling.
The result is stored as its own admin_save run with the real request method, separate from your site-wide numbers.
Open a post, page, product, CPT item or HPOS order for editing, make your edits, then use Analyse update/save in the PA admin-bar menu instead of the Update button.
The control only appears on those screens. From the shared Super Speedy dashboard, "Analyse a post or product save" opens the most recent published item with the tool primed.
Start with the phase table. A slow save is nearly always one of three things, and the phases separate them: plugin boot callbacks (something expensive loads on every admin request), the write itself (save_post and friends), or outbound HTTP during the save (a sync to an external service running inline instead of in the background).
Mail time is broken out. A save that sends notification email is paying for your mail server inside the request.
An existing item. There's no measurement for creating a brand new one, because the profiled request has to be a real save of real edits.
Administrator only, on your own site, same-origin.
Related: Per-Page-Analysis · Plugin-Impact-Analysis · How-It-Measures
Super Speedy Performance Analysis is free and GPLv3. Download the latest release · Knowledge base · Report a problem
Start here
What it can show you
- Per-Page-Analysis
- Checkout-and-Order-Flow-Analysis
- Update-and-Save-Analysis
- Plugin-Impact-Analysis
- Cache-Optimisation-Analysis
- The-Community-Plugin-Database
Getting the full picture
Reference
Project