Prometheus is an open source monitoring tool that works in a pull fashion: instead of each individual application sending its collected metrics to the server, each application exposes an endpoint and the Prometheus server constantly scrapes the available metrics. It is also able to dynamically discover metric sources to scrape, and it comes with support out of the box for different sources, such as AWS EC2, Azure VM, and Google Container Engine instances.
On top of this, Prometheus provides a powerful query language and is also able to send alerts based on the collected data. These are the types of metrics Prometheus is able to collect:
- Counter: This allows you to count whenever something happens. By definition, counters can only increase.
- Gauge: This lets you track quantities that vary over time, such as CPU time or number of Elixir processes running in the BEAM. Each data point captures the value of the gauge metric at a specific moment in time.
- Histogram: This lets you...