feat(metrics): bucket md viewer edit volume into LTE500 / LTE1K / GT1K

Adds a session-cumulative edit-batch counter in _onIframeContentChanged
that fires one-shot count events as the user crosses each bucket
threshold:

  md/edits/LTE500   first time a session reaches >=500   edit batches
  md/edits/LTE1K    first time a session reaches >=1000  edit batches
  md/edits/GT1K     first time a session crosses 1001+   edit batches

Each bucket fires at most once per session, so the funnel reads as
LTE500 >= LTE1K >= GT1K — answering "what fraction of users edit
heavily in the markdown live preview?"

The counter only increments on iframe content changes (md viewer edit
mode), not on direct CodeMirror edits, since CM bypasses
_onIframeContentChanged. A "batch" is a 50ms-debounced burst of edits
(set by the iframe's existing debounce), so the buckets are
representative volumes rather than per-keystroke counts.

Author: abose

src/extensionsIntegrated/Phoenix-live-preview/MarkdownSync.js

@@ -53,13 +53,13 @@ define(function (require, exports, module) {
     // heavy markdown editors. Each iframe content change (already
     // 50ms-debounced upstream) increments this. When the count crosses
     // a bucket threshold we fire a one-shot count event so the analytics
-    // funnel reads as: users-with-LTE5 ⊇ LTE25 ⊇ LTE100 ⊇ GT500.
+    // funnel reads as: users-with-LTE500 ⊇ LTE1K ⊇ GT1K. A representative
+    // sample, not a fine-grained histogram.
     let _mdEditCount = 0;
     const MD_EDIT_BUCKETS = [
-        { threshold:   5, label: "LTE5"   },
-        { threshold:  25, label: "LTE25"  },
-        { threshold: 100, label: "LTE100" },
-        { threshold: 500, label: "GT500"  }
+        { threshold:  500, label: "LTE500" },
+        { threshold: 1000, label: "LTE1K"  },
+        { threshold: 1001, label: "GT1K"   }
     ];
 
     let _active = false;