CLI: Add site editor performance benchmark

[brandonpayton]

Summary

• Add a Playwright-based site editor performance benchmark for the Playground CLI, adapted from Studio's tools/benchmark-site-editor/
• Measure 6 metrics: server startup time plus 5 site editor interaction metrics
• Support both unbuilt-jspi and built CLI modes via --mode flag, with configurable rounds and an optional plugins-loaded variant

How it works

The benchmark spawns a Playground CLI server via Nx targets (npx nx unbuilt-jspi playground-cli or npx nx start playground-cli), then launches headless Chromium to navigate through the WordPress site editor and time each interaction. Results are aggregated using median across rounds and output as both a console table and a JSON artifact.

Metrics measured:

Metric	Description
serverStartup	Time from spawning the CLI process until the server responds to HTTP requests
siteEditorLoad	Time from clicking Appearance → Editor until blocks render in the editor canvas
templatesViewLoad	Time to open the Templates view and load the template grid
templateOpen	Time to open a specific template in the editor
blockAdd	Time to insert Paragraph + Heading blocks via the Block Inserter
templateSave	Time to save the template until "Saved" button appears

Usage:

npx nx perf playground-cli
npx nx perf playground-cli -- --mode=built --rounds=5 --with-plugins
npx nx perf playground-cli -- --headed --rounds=1

Test plan

• Run npx nx perf playground-cli -- --rounds=1 and verify it starts the CLI, runs measurements, prints a results table (including serverStartup), and saves a JSON artifact
• Run with --mode=built and verify it builds first, then benchmarks the built CLI
• Run with --with-plugins and verify both bare and with-plugins environments are benchmarked
• Run with --headed and verify Chromium launches visibly for debugging
• Verify npx nx lint playground-cli passes (no no-console errors from perf files)

🤖 Generated with Claude Code

[Copilot]

Pull request overview

Adds a Playwright-driven performance benchmark workflow for the Playground CLI that measures site editor interaction timings and persists results as JSON artifacts.

Changes:

• Introduces a new nx perf playground-cli target and a root npm script to run the benchmark.
• Adds a Playwright measurement harness for 5 site editor interaction metrics, plus server startup timing.
• Adds a “with plugins” blueprint and artifacts output ignore rules for repeatable runs.

Reviewed changes

Copilot reviewed 7 out of 8 changed files in this pull request and generated 6 comments.

Show a summary per file

File	Description
packages/playground/cli/project.json	Adds an Nx perf target to run the benchmark entrypoint.
packages/playground/cli/perf/benchmark.ts	Implements server spawn/teardown, rounds/median aggregation, and JSON/table reporting.
packages/playground/cli/perf/measure-site-editor.ts	Implements Playwright steps for site editor navigation and metric collection.
packages/playground/cli/perf/plugins-blueprint.json	Adds a plugin-heavy blueprint variant for benchmarking.
packages/playground/cli/perf/artifacts/.gitignore	Ensures generated benchmark artifacts aren’t committed.
packages/playground/cli/.eslintrc.json	Disables no-console for perf scripts.
package.json	Adds a convenience npm script to run the benchmark.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[brandonpayton]

This is pretty separate from everything else. It works when I run it locally, and all the existing tests pass.

Now we can use this for better evaluation of specific Windows performance improvements. If we want, we could also post performance output for every PR that touches a Playground CLI dependency.

Let's merge once the tests pass again after a small update.

[brandonpayton]

note: I meant to commit with a [CLI] commit message prefix, but committed with a CLI: prefix.