One external monitor never comes back after the idle screen-off timeout
Strong record with one API defect and one stale citation. Verified: hl.dsp.dpms({ action?, monitor? }) and hl.dsp.force_renderer_reload() are both documented, the Monitor param accepts a name or desc:, and hyprctl dispatch 'hl.dsp.dpms({ action = "enable" })' is exactly the form Omarchy uses in omarchy-brightness-display and omarchy-hyprland-monitor-clamshell. misc:mouse_move_enables_dpms and misc:key_press_enables_dpms both default to false as stated. hypridle.conf is still hyprlang/ini, so "do not convert it to Lua" is right. #2437 is real and its body does contain [ERR] Couldn't commit output DP-5 on a Thunderbolt-attached Dell. Discussion #11230 is real and matches its summary. wlopm is in extra, so pacman -S --needed wlopm works (the yay fallback is unnecessary but harmless). The Omarchy 4 section is exact: hypridle is not in Quattro at all, ~/.config/omarchy/shell.json carries a top-level idle block with screensaver and lock in seconds (the record's JSON matches manual/13-toggles-idle-screensaver.md verbatim, defaults included), and omarchy-debug-idle exists. The defect: hl.set("misc:mouse_move_enables_dpms", true) is not a real API. Only hl.config({ category = { option = ... } }) is documented, and hl.set appears nowhere in the wiki or the Omarchy tree. Step 1 is the first thing a reader pastes, and it takes the config file down with it. Separately, #2437 is closed as *completed* and was resolved via the wlroots WLR_DRM_NO_MODIFIERS workaround. Hyprland has used Aquamarine, not wlroots, since 0.42, so presenting it as the live mechanism is stale. #11230 and #14234 are the current threads.
The Hyprland wiki warns that binding DPMS directly to a key "might cause undefined behavior". Call it from a timer or a script, not straight from a bind. If you experiment with dpms disable on your only monitor and input wake is not configured, you will have a black but running session: switch to a TTY with Ctrl+Alt+F2 and run hyprctl dispatch 'hl.dsp.dpms({ action = "enable" })' with HYPRLAND_INSTANCE_SIGNATURE exported, rather than power-cycling.
Symptom
hypridle blanks the screens after the idle timeout. Moving the mouse brings back the laptop panel and one external monitor, but the third display stays completely black. The only way back is to unplug and replug it, or reboot. The journal shows Couldn't commit output DP-5. Sometimes the reverse happens on the way down: hyprctl dispatch dpms off from hypridle turns off only the primary monitor while the others stay lit, then flicker when I come back.
Cause
Hyprland's DPMS dispatcher issues a DRM-level power change per output, and on some links (Thunderbolt/USB4 docks, DisplayPort MST, NVIDIA) the commit for one output fails and is never retried, leaving that screen dark while the others come back. The all-monitors form of the dispatcher is also unreliable when driven by hypridle rather than by hand. Discussion #11230 reports that a manual invocation turns off all monitors while the hypridle-triggered one regularly misses some, and #14234 tracks monitors that will not come back on at all. (Historical note: #2437 is the frequently-cited report of Couldn't commit output DP-5 on a Thunderbolt-attached Dell, but it is closed as completed and was traced to a wlroots bug fixed with WLR_DRM_NO_MODIFIERS. Hyprland has run on Aquamarine rather than wlroots since 0.42, so that specific fix no longer applies and the current failures are a different code path with the same symptom.) On top of the commit failures, misc:mouse_move_enables_dpms and misc:key_press_enables_dpms both default to false, so input does not wake the outputs at all unless something explicitly re-enables them.
Fix
Steps 2 through 5 and the Omarchy 4 section are correct as written. Keep the per-output hypr-dpms-on retry loop, the monitors all filter, the manual hl.dsp.dpms({ action = "enable", monitor = "DP-5" }) rescue, hl.dsp.force_renderer_reload(), and the wlopm alternative. Only step 1 needs replacing.
**1. Let input wake the displays.** Both options default to false. The documented Lua syntax is hl.config, not hl.set. Add to ~/.config/hypr/looknfeel.lua:
hl.config({
misc = {
mouse_move_enables_dpms = true,
key_press_enables_dpms = true,
},
})
hyprctl reload
To test the pair before committing them to a file:
hyprctl eval 'hl.config({ misc = { mouse_move_enables_dpms = true, key_press_enables_dpms = true } })'
One addition to step 3: give hypr-dpms-on a wlopm --on '*' last resort at the end of the retry loop, since it goes through wlr-output-power-management-v1 rather than Hyprland's DRM path and often succeeds where the dispatcher has already failed:
command -v wlopm >/dev/null && wlopm --on '*' 2>/dev/null || true
Verify
Trigger the timeout (or run the disable dispatcher by hand), then move the mouse. Every output should light up. hyprctl monitors all -j | jq -r '.[] | "\(.name) \(.width)x\(.height) dpms=\(.dpmsStatus)"' should report dpmsStatus: true and a non-zero mode for all of them.
Sources
- https://github.com/hyprwm/Hyprland/issues/2437
- https://github.com/hyprwm/Hyprland/discussions/11230
- https://github.com/hyprwm/Hyprland/discussions/14234
- https://raw.githubusercontent.com/hyprwm/hyprland-wiki/main/content/Hypr%20Ecosystem/hypridle.md
- https://raw.githubusercontent.com/hyprwm/hyprland-wiki/main/content/Configuring/Basics/Dispatchers.md
- https://raw.githubusercontent.com/hyprwm/hyprland-wiki/main/content/Configuring/Basics/Variables.md
- https://raw.githubusercontent.com/basecamp/omarchy/quattro/bin/omarchy-debug-idle
- https://raw.githubusercontent.com/basecamp/omarchy/quattro/manual/13-toggles-idle-screensaver.md