v0.35.3

CHANGELOG.md

@@ -2,12 +2,18 @@
 
 Release history focused on **features, settings, default behavior, compatibility and fixes**. Changes to internal code structure and preliminary performance experiments are intentionally omitted. Historical descriptions reflect their respective releases; consult the current documentation for the settings available in the latest release.
 
+## 0.35.3 — Overview layout fixes
+
+- **Standard Overview layout restored:** the quick-health tiles added in 0.35.2 (healthy backends, error share, requests rejected by limits, nearest certificate expiry) are now hidden by default and placed at the end of the Overview, so they no longer shift the standard block layout. Enable them in the layout editor when needed.
+- **Fixed:** the "Saved default" status next to the history sample count was always visible and looked like a fresh save notification. It now appears only for an unsaved temporary choice or briefly after **Remember choice**.
+- **Upgrade notes:** no configuration changes are required. If 0.35.2 already changed your Overview, use **Unlock layout → Reset layout** to restore the default arrangement, or hide the tiles in the layout editor.
+
 ## 0.35.2 — memory fix, health check options and Overview tiles
 
 - **Memory use restored:** with a large HOT cache, 0.35.1 used about 25% more memory than 0.35.0 (for example 1.1–1.2 GiB instead of about 800 MiB with a 600 MiB cache). The default build is back to the 0.35.0 memory profile. The faster alternative allocator is available as an optional build feature for servers with spare RAM; the Performance guide explains the trade-off with measurements.
 - **Health check port:** active TCP and HTTP health checks can probe a different port than the one receiving traffic, for example a status port 8079 for a backend serving on port 80. It can be set for a whole Vhost or per backend; host, TLS and address restrictions stay the same.
 - **Expected response text:** HTTP health checks can require a text fragment in the response body, so a status page that answers 200 while reporting a degraded state marks the backend as down.
-- **New Overview tiles:** healthy backends (with the names of failing ones), error share, requests rejected by limits and the nearest certificate expiry are shown on the Overview by default.
+- **New Overview tiles:** healthy backends (with the names of failing ones), error share, requests rejected by limits and the nearest certificate expiry.
 - **Certificate expiry metrics:** Prometheus now exports the expiry date and remaining days of every installed certificate, including manually uploaded ones, for alerting.
 - **Safari:** the Overview no longer jumps when you scroll near the Backends table while live values refresh.
 - **Fixed:** network interface traffic below 1 KiB/s is shown as a whole number of bytes instead of a long decimal.