minerstat mining tutorial #3: Alerts and triggers

To stay aware of what is going on with your mining rigs, minerstat offers you several alerting options and custom triggers. Both are account-wide and are taken into consideration for each worker separately.

Alerts

Currently, there are 6 different types of notifications you can receive through 3 different methods. Each notification has a custom setting and each option is checked once a minute to see if an alert needs to be sent out.

GPU temperature
You get notified when any of your GPUs exceeds the temperature defined in settings: 60°C, 65°C, 70°C, 75°C, 80°C, 85°C, 90°C, or 95°C. Each GPU is checked for each worker separately.

When the worker goes offline
You get notified when the worker is offline for 4 minutes, 6 minutes, 8 minutes, 10 minutes, or 15 minutes.

When the worker comes back online
You get notified when the worker is back online for 4 minutes, 6 minutes, 8 minutes, 10 minutes, or 15 minutes.

ASIC notifications
You get notified when ASIC reports hardware errors and high temperature for 4 minutes, 6 minutes, 8 minutes, 10 minutes, or 15 minutes.

Hashrate drop
You get notified when the worker’s hashrate drops by 5%, 10%, 20%, 30%, 40%, or 50%. The last hashrate is compared to hashrate detected 10 minutes ago.

Efficiency rate drop
You get notified when the worker’s efficiency drops by 10%.

The 3 methods through which you can receive notifications.

E-mail
You get notified to the e-mail address you used for registration on minerstat.

Telegram
You get notified to your Telegram when you connect with Telegram. Click on the green button in your alerts settings and confirm with our minerstat bot. After you make connections, your Telegram notifications are ready. You can switch the Telegram ID at any time you want.

Mobile device
You get notified to your mobile through push notifications after you connect your mobile device with minerstat:

What else you need to know about the alerts?

Events displayed in worker’s stats and diagnostics are not connected to the alerts. While we monitor your workers for custom alerts every minute, the diagnostics only show historical data based on 10-minutes (or 30-minutes if you are a free user) statistics data. If your worker goes offline for 4 minutes and comes back after 2 minutes and this all happens between 10-minute checking period, historical data won’t detect this event, but the alerting system will and will notify you about it.

If your client restarts (you can see it as a shorter last sync time than expected) and after few seconds client is back to mining, we won’t alert you about it. Your worker must be offline for the minimum of 4 minutes to receive an alert. Similarly, this won’t get detected in historical data as it is only checked every 10 minutes. If your client is restarting a lot, we suggest to inspect it and find the reason for it. You can even join our Discord channel and discuss it with us and/or other miners.

Triggers

The second, more advanced feature connected to different events are triggers, which help you with automatic actions when one or more events are detected. You can select between 6 events with different settings, which can trigger 5 different actions.

GPU temperature
A custom trigger can be fired when GPU temperature is higher or equal than 60°C, 65°C, 70°C, 75°C, 80°C, 85°C, 90°C, or 95°C.

ASIC temperature
A custom trigger can be fired when ASIC temperature is higher or equal than 60°C, 65°C, 70°C, 75°C, 80°C, 85°C, 90°C, or 95°C.

Hashrate drop
A custom trigger can be fired when hashrate is down by 5%, 10%, 20%, 30%, 40%, or 50%. The last hashrate is compared to the average hashrate in the last 60 minutes.

Efficiency rate drop
A custom trigger can be fired when efficiency is smaller or equal to 90%.

Time
A custom trigger can be fired when the time is any full hour (from 00:00 to 23:00).

Unresponsive
A custom trigger can be fired when your rig is unresponsive for 20, 30, 40, 50, or 60 minutes.

Idle
A custom trigger can be fired when your rig is idle for 5, 10, 15, 20, 25, 30, 35, 40, 45, 50, 55, or 60 minutes. Idle means that your msOS is online, but your mining client isn’t. In this case you can either reboot or shut down the machine.

The actions that can be triggered by events are the following.

Restart software
This action will restart the mining clients and software on the rig on which the action applies.

Reboot machine
This action will turn off clients and reboot the mining rig on which the action applies.

Shut down machine
This action will turn off clients and shut down the mining rig on which the action applies. Be extra careful with action since you will have to be physically available at the location of your machine to turn it back on.

Set fans
This action will set the fans on the mining rig for which the action applies to the selected value. You can choose any value between 50% and 100% fan speed. Be extra careful with this action since it will overwrite your overclock settings for the rig(s) on which the actions apply.

Send email
This action can be chosen only when your rig is unresponsive. This combination is practical when you have an external monitoring system intact which cannot communicate with the outside world if the internet is down, the rig freezes, etc., but they can still receive e-mails or read API responses (for example IFTTT).

It is important to know that the triggers will always check the most extreme option first. For example, let’s say that you have set the following two triggers:

If your GPU temperature will all of the sudden reach 90°C, triggers will restart your software without setting the fans to 80%. The fans will be set to 80% if detected GPU temperature will be between 75°C and 85°C.

If you have any questions or suggestions for alerting system and triggers, join us on our Discord channel.

More from this series: