uptime-monitor

Uptime monitor

This is a set of scripts for monitoring machine crashes. Run the client on your vast machine and the server on a remote one. You get notifications on Telegram if no heartbeats are sent within the timeout (default 12 seconds).

I recommend a $2.50 or $3.50 server from Vultr. Server location must be set to New Jersey for the cheap plans. Use my referral link for $100 credit. https://www.vultr.com/?ref=8581277-6G

Setup Telegram bot

1. Search for the contact BotFather in app
2. Send /newbot as a message
3. Enter the details and then copy the token it gives you
4. Create a group chat with your bot and send "/start"

Config setup

Create a file called ".env" in this directory. You can copy the same .env between server and client, only some are absolutely required though.

CHAT_ID=                  # SERVER only - See below for steps
TELEGRAM_TOKEN=<token>    # SERVER - Token as given in previous step
FAIL_TIMEOUT=5            # SERVER - If no pings in 5 seconds, send alert.
PING_INTERVAL=2           # CLIENT - How often to send pings
API_KEY=asecretkey        # SERVER+CLIENT - A random unique key to authenticate the client
SERVER_ADDR=192.168.1.150 # SERVER+CLIENT - Address the client will use to send pings
SERVER_PORT=5000          # SERVER+CLIENT - Port of the server.

If you get too many false notifications, try increase the FAIL_TIMEOUT.

Server setup

Install dependencies

pip install -r requirements.txt

Get chat id

After adding your telegram_token to your .env file and sending /start to your bot, run the below to get the chat_id and also add it to the .env file.

./run_server.sh chat_id

Run server

./run_server.sh

Start on boot (optional) change path to uptime-monitor location

(crontab -l; echo "@reboot screen -dmS uptime-server /root/uptime-monitor/run_server.sh") | crontab -

Client setup

Run client. Pass a name (no spaces) for your worker.

./run_client.sh <worker_name>

Start on boot (optional) change path and <worker_name>

(crontab -l; echo "@reboot screen -dmS uptime-client /root/uptime-monitor/run_client.sh <worker_name>") | crontab -