Connection Watchdog | Docs

Driver name in Composer: Connection Watchdog. Manufacturer: Theater Monster. File: connection_watchdog.c4z. Current driver version: 2026110803 (versions are date stamps). Store slug: connection-watchdog.

Overview

Connection Watchdog watches every network connection in the project using Control4's own online status (what Composer's Network tab shows). When a device that has been online drops and stays down past a grace period, it fires Device Offline. It can then try a soft reconnect, Wake-on-LAN, and finally power-cycle the device's WattBox outlet or Araknis PoE port. If the device is still down after the recovery wait, it fires Recovery Failed.

Also:

Tells "device is gone" from "driver lost it" with a short network probe (shown in Last Diagnosis).
Network Outage detection: when many devices drop at once, it holds per-device recovery (the cause is a switch, router or power).
Device Address Changed event when DHCP moves a device.
Unstable Device event for devices that keep dropping.
Daily uptime summary.
Optional dealer dashboard: posts every event plus a heartbeat every 15 minutes.
Power-aware: a TV or device that is turned off is "asleep", not offline.

Who it is for: dealers who want to know a device dropped before the customer calls, and fix common drops without a truck visit.

Requirements

Control4 OS 3.0.0 or newer.
No hardware. For automatic power cycling: a WattBox outlet or Araknis PoE port already in the project as a relay, outlet or light device.
For the dealer dashboard: controller internet access, plus a dashboard URL and alert key.
Watches IP devices only. RS-232 and IR devices are not visible to it.

Installing and first setup

Add Connection Watchdog to any room. The first check runs 30 seconds after Director starts. Every IP device is watched from then on.
Recovery slots (optional): set Recovery Slots to how many devices need a recovery action (1 to 64, default 6). Slots only add recovery; every device is watched either way. Lowering the number hides slots but keeps their settings.
For each slot:

Slot N Device: pick the device.
Slot N Power: pick the relay for its WattBox outlet or Araknis PoE port.
Slot N Reconnect Command (optional softer step): the device driver's own command name, exactly as in that driver's Commands list (for example Reconnect).
Slot N MAC (optional): for a TV that sleeps in network standby, to send Wake-on-LAN.

Run Test Reconnect and Test Power Cycle for each slot to confirm the right device and outlet.
Set Auto Power Cycle to On.
Programming: on Device Offline or Recovery Failed, send a push notification with variable LAST_OFFLINE_DEVICE.
Dealer dashboard (optional): set Site Name, Dealer Alert URL and Dealer Alert Key, then run Send Test Alert. Dealer Alert Status shows the result.

Properties

Status

Driver Version.
Status: current monitoring state, or Monitoring off.
Offline Devices: devices currently offline.
Last Diagnosis: last probe result with time, for example "device is gone" or "device is up, driver lost it", and when it recovered.
Daily Summary: devices below 99% uptime yesterday.
Dealer Alert Status: Not configured, OK with time, or Failed with the reason and how many alerts are queued.

Monitoring settings

Site Name: the customer name shown on the dealer dashboard.
Dealer Alert URL / Dealer Alert Key: dashboard address and key. Leave blank to turn the dashboard off.
Monitoring: On (default) or Off.
Check Interval (sec): 15 to 600, default 60.
Offline Grace (min): 1 to 60, default 3. How long a device must stay down before Device Offline.
Idle Driver Grace (min): 5 to 240, default 30. If an offline device still answers on its port, its driver is just idle (many drivers connect only when needed). No alert until this time passes, and it is never power-cycled.
Monitor Zigbee: Off (default) or On.
Network Probe: On (default) or Off. The short connect test used to tell a dead device from an idle driver. The watchdog only probes devices that are already offline.
Outage Threshold (%): 10 to 100, default 30. With 3 or more devices down and at least this share of devices that have been online, it declares a Network Outage.
Unstable After Drops (24h): 2 to 20, default 4. Only drops of a minute or longer count, and not during an outage.
Ignore Device IDs: device IDs to skip, separated by commas or spaces.
Debug Mode: Off (default) or On.

Recovery settings

Reconnect Wait (min): 1 to 30, default 2. Time after the reconnect command or Wake-on-LAN before moving on to a power cycle.
Auto Power Cycle: Off (default) or On.
Power Off Time (sec): 5 to 120, default 15. How long power stays off.
Recovery Wait (min): 2 to 60, default 5. Time after a power cycle before Recovery Failed.
Max Power Cycles Per Day: 1 to 10, default 3, per device. Prevents boot loops.
Recovery Slots: 1 to 64, default 6.
Slot N Device, Slot N Power, Slot N Reconnect Command, Slot N MAC: see setup.

Actions

Check Now: runs a connection check now.
Refresh Device List: re-reads the project's network connections and refreshes the Slot device lists.
Print Status: prints every monitored connection to the Lua tab.
Send Test Alert: posts a test alert to the dealer dashboard.
Test Power Cycle (Slot): power-cycles that slot's outlet or PoE port now.
Test Reconnect (Slot): sends that slot's reconnect command and/or Wake-on-LAN now.

Programming

Events:

Device Offline: a device that had been online stayed offline past the grace period (LAST_OFFLINE_DEVICE).
Device Recovered: it is back online (LAST_RECOVERED_DEVICE).
Reconnect Attempted: the slot's reconnect command and/or Wake-on-LAN was sent (LAST_RECONNECT_DEVICE).
Power Cycle Performed: an outlet or PoE port was cycled (LAST_POWER_CYCLED_DEVICE).
Recovery Failed: still offline after the power cycle. Needs a service call. Fires only for devices with a recovery slot.
Device Address Changed: a device's IP changed, likely DHCP (LAST_ADDRESS_CHANGE).
Network Outage / Network Restored: many devices dropped at once (per-device recovery on hold), then cleared.
Unstable Device: dropped repeatedly in 24 hours (LAST_UNSTABLE_DEVICE).
Daily Summary: yesterday's uptime summary is ready (DAILY_SUMMARY).

Variables: OFFLINE_COUNT, LAST_OFFLINE_DEVICE, LAST_OFFLINE_DEVICE_ID, LAST_RECOVERED_DEVICE, LAST_POWER_CYCLED_DEVICE, LAST_ADDRESS_CHANGE, LAST_RECONNECT_DEVICE, LAST_UNSTABLE_DEVICE, DAILY_SUMMARY, NETWORK_OUTAGE.

Commands for programming: Check Now, Refresh Device List, Send Test Alert, Print Status.

No conditionals.

Connections and bindings

None. The watchdog reads the project's network connections directly. Power devices are chosen in the Slot N Power properties, not bound.

Best practices

Set up slots only for devices that can safely be power-cycled (streamers, gateways, switches on a PoE port). Leave TVs to Wake-on-LAN.
Test every slot with Test Power Cycle before turning on Auto Power Cycle.
Act on Device Address Changed by setting a DHCP reservation on the router.
Act on Unstable Device by checking the cable, switch port or the device's network card.
Add devices with known noisy drivers to Ignore Device IDs.
Daily power-cycle counts, uptime and queued alerts survive a Director restart.
If the site's internet is down, dashboard events are queued (up to 100) and sent later. The dashboard flags a site that stops checking in.

Troubleshooting

A device never shows as offline: it must have been seen online first. HTTP-only drivers that always show yellow in Network Tools are ignored. Serial and IR devices are not visible.
TV reported offline at night: should not happen on current versions; a device whose power state is off is treated as asleep. Update to 2026110802 or newer.
On-demand drivers (for example the Samsung TV by Papaya) cause offline alerts: since 2026110802 the watchdog probes first. A device that answers is an idle driver; no alert until Idle Driver Grace.
Status frozen and no dashboard posts on a project with a TV that has two network ports: fixed in 2026110803 (devices with several ports were crashing the check). Update the driver.
Recovery Failed never fires: it only fires for devices with a reconnect command, MAC or power device in a slot.
Power cycle hit the wrong outlet: check Slot N Power and use Test Power Cycle.
Reconnect does nothing: Slot N Reconnect Command must match the device driver's command name exactly.
Dealer Alert Status shows Failed: check Dealer Alert URL, Dealer Alert Key and the controller's internet. A failure that says no key sent or wrong key means the key is missing or wrong.
Recovery held during an outage: expected. When the outage clears, Network Restored fires and any device still down gets a fresh grace period.
Composer says "Unable to add to <room>" with a blank name: the driver file was renamed. The file must stay connection_watchdog.c4z.

Limitations

IP devices only. No RS-232 or IR monitoring.
Cannot change another device's IP address. DHCP changes are detected and reported; the fix is a reservation or an SDDP binding.
Cannot keep another driver's socket alive. It never opens a connection to an online device.
Power cycling needs a WattBox outlet, Araknis PoE port or similar relay/outlet device in the project.
Wake-on-LAN depends on the device supporting it.

Version history highlights

2026110803: fixed a crash when one device has two network ports (one online, one offline).
2026110802: power-aware (devices turned off are asleep), probe before alerting with Idle Driver Grace, Unstable Device counts only real drops. Encrypted.
2026110801: driver code encrypted.
2026102906: fixed dashboard posts failing after a driver reload.
2026102905: Last Diagnosis shows recovery time; Recovery Failed only for devices with a recovery path.
2026102903 to 2026102904: outage guard, reconnect command and Wake-on-LAN step, dealer dashboard, Unstable Device, daily summary, persisted state.
October 2026: first release.