This module provides remote monitoring of a running process over HTTP. It can be used to run an HTTP server that provides both a web-based user interface and a machine-readable API (e.g. JSON.) The former can be used by a human to get an overview of what the program is doing and the latter can be used by automated monitoring tools.

Typical usage is to start the monitoring server at program startup

main = do
    forkServer "localhost" 8000
    ...

and then periodically check the stats using a web browser or a command line tool (e.g. curl)

$ curl -H "Accept: application/json" http://localhost:8000/

To make full use out of this module you must first enable GC statistics collection in the run-time system. To enable GC statistics collection, either run your program with

+RTS -T

or compile it with

-with-rtsopts=-T

The runtime overhead of -T is very small so it's safe to always leave it enabled.

Be aware that if the server started by forkServer is not bound to "localhost" (or equivalent) anyone on the network can access the monitoring server. Either make sure the network is secure or bind the server to "localhost".

To use the machine-readable REST API, send an HTTP GET request to the host and port passed to forkServer.

The API is versioned to allow for API evolution. This document is for version 1. To ensure you're using this version, append ?v=1 to your resource URLs. Omitting the version number will give you the latest version of the API.

The following resources (i.e. URLs) are available:

/

JSON object containing all metrics. Metrics are stored as nested objects, with one new object layer per "." in the metric name (see example below.) Content types: "text/html" (default), "application/json"

/<namespace>/<metric>

JSON object for a single metric. The metric name is created by converting all "/" to ".". Example: "/foo/bar" corresponds to the metric "foo.bar". Content types: "application/json"

Each metric is returned as an object containing a type field. Available types are:

• "c" - Counter
• "g" - Gauge
• "l" - Label
• "d" - Distribution

In addition to the type field, there are metric specific fields:

• Counters, gauges, and labels: the val field contains the actual value (i.e. an integer or a string).
• Distributions: the mean, variance, count, sum, min, and max fields contain their statistical equivalents.

Example of a response containing the metrics "myapp.visitors" and "myapp.args":

{
  "myapp": {
    "visitors": {
      "val": 10,
      "type": "c"
    },
    "args": {
      "val": "--a-flag",
      "type": "l"
    }
  }
}

The monitoring server can store and serve integer-valued counters and gauges, string-valued labels, and statistical distributions. A counter is a monotonically increasing value (e.g. TCP connections established since program start.) A gauge is a variable value (e.g. the current number of concurrent connections.) A label is a free-form string value (e.g. exporting the command line arguments or host name.) A distribution is a statistic summary of events (e.g. processing time per request.) Each metric is associated with a name, which is used when it is displayed in the UI or returned in a JSON object.

Metrics share the same namespace so it's not possible to create e.g. a counter and a gauge with the same. Attempting to do so will result in an error.

To create and use a counter, simply call getCounter to create it and then call e.g. inc or add to modify its value. Example:

main = do
    handle <- forkServer "localhost" 8000
    counter <- getCounter "iterations" handle
    let loop n = do
            inc counter
            loop
    loop

To create a gauge, use getGauge instead of getCounter and then call e.g. set. Similar for the other metric types.

It's also possible to register metrics directly using the System.Metrics module in the ekg-core package. This gives you a bit more control over how metric values are retrieved.