Upbeat is a high performance node-based healthcheck/dashboard. Upbeat allows you to run health checks and provides a dashboard to chart the performance. It also allows you to proxy and cache these health checks so they don't tax your system. You can now reduces the number of health checks to a service from O(N) to O(1).
npm install -g upbeat
npm install upbeat
var upbeat = require('upbeat'); var server = new upbeat.Server(configObject); server.run();
dashboard: port: 2468 sync: redis: port: 6379 host: localhost services: google: www: strategy: http url: http://www.google.com connection: strategy: tcp host: google.com port: 80
Services are a way of grouping several sensor checks together. In the example above, we have a "google" service w hich we check by making a get request to "http://www.google.com" and also seeing if we have a connection to port 8 0 on the "google.com" host. In the yaml config, a service is a "hash" of sensors where the keys are the names of the sensors and the values are the configuration.
Sensors are a way of describing a health check. Each sensor config MUST at least have a strategy. Common configuration options accross all strategies are:
- timeout: number (milliseconds) to define how long it will allow a check before declaring it a failure
- interval: number (milliseconds) of time to wait between health checks (called after the result of a check)
- fall: number of fails to be considered down
- rise: number of passes to be considered up
Here are some examples of how you can use sensors and their strategies:
strategy: tcp host: google.com port: 80
strategy: http url: http://www.google.com
strategy: pidfile pidfile: /var/pids/mysql.pid
strategy: redis host: localhost # defaults to localhost port: 6379 # defaults to 6379 command: [ 'exists', 'foo' ]
strategy: mysql host: host # required user: user # required password: pass # required sql: "SELECT * FROM users LIMIT 1" # defaults to "SHOW TABLES"