Watchdog
A watchdog is a guard that restarts a frozen ASIC without you. The firmware kind watches the mining process, the external kind watches the ASIC itself.
The kinds of watchdogs
The firmware watchdog lives inside the ASIC: if the mining process hangs, stops reporting hashrate or loses hashboards, the guard restarts the process or the whole ASIC. In custom firmware it is configurable, in stock firmware it is simply there.
An external watchdog is a device between the outlet and the ASIC: a smart plug or relay that pings the ASIC or reads its hashrate over the API. No answer for a few minutes, and the power gets cycled. It saves the day where the firmware kind is helpless: when the control board itself freezes solid.
Scheduled reboots, once a day at night, exist too. That is not a watchdog, it is prophylaxis, and opinions differ: a healthy ASIC doesn't need it, a troubled one gets to limp along until repair.
The general rule: a watchdog treats the symptom. An ASIC that freezes every day is faulty, and the guard merely masks that with a clean uptime line. Frequent triggers are a reason to hunt for the cause, not to admire the automation.
The short version
| Firmware kind | restarts the process or the ASIC from inside |
| External kind | cycles power on ping or API silence |
| When you need external | the control board itself freezes |
| What is not a watchdog | a reboot on a schedule |
| Warning sign | triggers more than once a week |
How to line up your guards
Make sure the firmware watchdog is on: in custom firmware it sits in the settings, and you leave it off only while debugging.
At a remote site add an external one: a smart plug with ping costs about one trip to the site and pays for itself with the first freeze.
Count the triggers through monitoring: a guard without statistics hides the ASIC's decline.
Don't set aggressive thresholds: a restart over a one-minute network hiccup loses more than it saves.
An example from practice
An ASIC at a remote site froze solid every couple of weeks, and only a power cycle helped. A smart plug with ping closed the issue: downtime dropped from a day to five minutes. A month later the plug's statistics showed the freezes climbing to three times a week, and the ASIC went in for repair with a rotting control board: the guard bought time and delivered the diagnosis.
Related terms
Where to go next on the site
Software
- Software firmware
Hardware
- ASIC Miner knowledge base
Questions and answers
How is a watchdog different from monitoring
Monitoring tells you, a watchdog acts on its own. They work as a pair: the guard restarts, monitoring counts how often that happens.
Are frequent restarts harmful
By themselves no, an ASIC is built for them. The cause is what hurts: frequent freezes are a fault the guard is masking.
Is a scheduled reboot enough
No, a freeze doesn't check the schedule. A nightly reboot is prophylaxis, the guard is there for the incident, not the calendar.
How the terms connect
Every link in the chain is clickable. Orange marks where you are now.
Looking for an ASIC miner
The catalog holds 212 models. You can compare them by hashrate and by joules per terahash, then plug your own rate into the calculator and see what stays in your pocket.
Page written and checked by Denys Klimchuk. Updated .