A one-file NodeJS server status monitoring/notification tool.
Visit https://apulse.ybouane.com for a demo!
- Highly and easily configurable, copy config.js from config.dist.js, then edit file to add test endpoints and configure the watcher
- Supports sending outage notifications by: Telegram, Discord, Slack, SMS (Twilio API), Email (SendGrid API)
- Uses the Fetch API to test server-responses, you can configure GET, POST, PUT... requests and have full control over the fetch options.
- Check content for validity, HTTP status...
- Measures latency
- Minimal and easy to use dashboard
- Easy to setup. Run the
entrypoint.shscript and open the static/index.html page to view the dashboard. - Auto-reload of the config file (no need to restart the watcher)
- No dependencies
aPulse can be configured through the "config.js" file and it looks like this
export default {
interval : 15, // Interval in minutes between each pulse
nDataPoints : 90, // Number of datapoints to display on the dashboard
responseTimeGood : 300, // In milliseconds, this and below will be green
responseTimeWarning : 600, // In milliseconds, above this will be red
timeout : 5000, // In milliseconds, requests will be aborted above this
verbose : true, // Whether or not to output pulse messages in the console
readableStatusJson : true, // Format status.json to be human readable
logsMaxDatapoints : 200, // Maximum datapoints history to keep (per endpoint)
statusFile : './static/status.json', // Path to status.json file
telegram : {}, // optional, tokens to send notifications through telegram
slack : {}, // optional, tokens to send notifications through slack
discord : {}, // optional, tokens to send notifications through discord
twilio : {}, // optional, tokens to send notifications through twilio (SMS)
sendgrid : {}, // optional, tokens to send notifications through sendgrid (Email)
consecutiveErrorsNotify : 1, // After how many consecutive Errors events should we send a notification
consecutiveHighLatencyNotify : 3, // After how many consecutive High latency events should we send a notification
sites : [ // List of sites to monitor
{
id : 'google', // optional
name : 'Google',
endpoints : [ // Each site is a bunch of endpoints that can be tested
{
id : 'homepage', // optional
name : 'Homepage', // optional
link : 'https://www.google.com', // optional, for notifications and dashboard only, [defaults to endpoint.url], can be disabled by setting it to false
url : 'https://www.google.com', // required
request : { // optional, fetch options
method: 'GET',
},
mustFind : 'Feeling Lucky', // optional, String | Array | Regex | Function | AsyncFunction
mustNotFind : /Page not found/i, // optional, String | Array | Regex | Function | AsyncFunction
customCheck : async (content, response)=>{return true;}, // optional, Function | AsyncFunction -> Run your own custom checks return false in case of errors
validStatus : [200], // optional, Which http status should be considered non errors [defaults to 200-299]
}
]
}
],
};Clone the repo:
git clone https://github.com/ybouane/aPulse.gitEither run the entrypoint.sh script directly (you need to keep it running in the background)
cd aPulse./entrypoint.shOr use a tool like PM2 (prefered method):
npm install pm2 -gStart watcher.json
pm2 start pm2.jsonConfigure pm2 to automatically start during startup
pm2 startupSave current pm2 processes list
pm2 saveThe watcher.js script only takes care of running the status checks and updates the file defined by statusFile (by default static/status.json). The entrypoint.sh script patches static/client.js with this path before starting watcher.js. If you want to view the final result, you simply need to serve the files in the static/ folder (or wherever your dashboard points to). You can use Nginx with a config like:
# Pulse
server {
root /var/www/apulse/static/;
index index.html;
server_name apulse.ybouane.com;
location /favicon.ico {
return 301 "/favicon.png";
}
listen 80;
}Or use any other tool to serve those files like the npm http-server package:
cd static
npx http-server -o ./