strong-supervisor/bin/sl-run.txt

supervisor and monitor for node.js applications

usage: %MAIN% [options] [app [app-options...]]

Run an app, allowing it to be profiled (using StrongOps) and supervised.

`app` can be a node file to run or a package directory. The default value is
".", the current working directory. Packages will be run by requiring the first
that is found of:
  1. javascript file mentioned in `scripts.start` of package.json
    *** NOTE: the script is not run and arguments are not preserved, only the
        path of the script is used, eg:
          `node --nodearg script.js --scriptarg` => 'script.js'
          `node bin/www` => `bin/www`
        The parser is simple, so options that accept arguments `--flag value`
        will cause problems.
  2. server.js
  3. app.js
  4. result of require(app)
    1. `main` property of app package.json
    2. `app`.js
    3. `app`/index.js


Options:
  -h,--help          Print this message and exit.
  -v,--version       Print runner version and exit.
  -l,--log FILE      Write supervisor and worker output to FILE
                       (defaults to "-", meaning log to stdout).
  --no-timestamp-workers
                     Disable timestamping of worker log lines by supervisor.
  --no-timestamp-supervisor
                     Disable timestamping of supervisor log messages.
  --no-log-decoration
                     Disable decorating supervisor/worker log messages with
                       cluster id/pid
  --syslog           Send supervisor and collected worker logs to syslog,
                       unsupported on Windows.
  --metrics BACKEND  Report metrics to custom backend. Implies `--profile`.
  -p,--pid FILE      Write supervisor's pid to FILE, failing if FILE already
                       has a valid pid in it (default is no pid file).
  --cluster N        Set the cluster size (default is 'cpu', but see below).
  --profile          Inject node instrumentation, the default.
  --no-profile       Do not inject node instrumentation.
  -C,--control CTL   Listen for control messages on CTL (default `runctl`).
  --no-control       Do not listen for control messages.

Log FILE is a path relative to the app's working directory if it is not
absolute. To create a log file per process, FILE supports simple substitutions
of %p for process ID and %w for worker ID.

Supported metrics backends are:

- `statsd://[<host>][:<port>]`
- `graphite://[<host>][:<port>]`
- `syslog:[?[application=<application>][&priority=<priority>]` (syslog is the
  Unix logging framework, it doesn't exist on Windows)
- `splunk://[<host>]:<port>`
- `log:[<file>]`
- `debug:[?pretty[=<true|false>]]`

It is possible to use multiple backends simultaneously.

Cluster size N is one of:

- A number of workers to run
- A string containing "cpu" to run a worker per CPU