Skip to content

Configuration

Before you can run opsbox, you must know how to configure it.

Opsbox runs on a flexible, modular arguments-based configuration system.

Each and every plugin can specify desired parameters, needed information, and other details that it wants from the user, which the user can set from a file, command line, or environment variable.

Where to Specify Configuration

Argument values can be specified in 4 places:

  • The default arguments file (~/.opsbox.json)
  • A file specified with the --config <path> flag
  • Double-dashed command line arguments and flags
  • Environment variables.

This is also the resolution order!

Both the default and specified configuration files consist of a JSON dictionary, where each key corresponds to an argument name and each value is the value you want that argument to have.

Specifying your pipeline

This is an argument that is core to how opsbox functions and works. It is highly recommended to take a look over this next section.

Opsbox uses a series of plugins specified at runtime to analyze different environments.

These plugins have their own arguments and required settings that will be looked for upon startup, and must specified as a pipeline in the following format:

input_1,input_2-assistant_1-assistant_2-output_1,output_2

Where:

  • The first arguments are the input plugins we want to use, seperated with a comma
  • The middle arguments are the assistants we want to use, seperated by a hyphen
  • The last arguments are the outputs we want to use, sepereated by a comma

For instance, if we wanted to get data related to AWS EBS volumes that have gone astray, we can use the stray_ebs plugin from our plugins repo and specify a pipeline like so:

stray_ebs-cost_savings-cli_out

In this pipeline:

  • We take stray_ebs as our input.
  • cost_savings will work on the data from stray_ebs and format it in it's own way (providing cost savings recommendations)
  • cli_out will handle the formatted output of cost_savings and display it to the user on the cli.

Note

Keep in mind plugins all have their own parameters they look for!

Logging

Opsbox provides a lot of customization over it's logging practices.

There are 4 configuration parameters that help you see what you want to see:

  • --log_level, which specifies the minimum level of logs you want to see. Must be one of "TRACE", "DEBUG", "INFO", "WARNING", "ERROR", and "CRITICAL".
  • --log_file, which specifies a path to a log file you want to output to.
  • --init_debug, which allows you to see traces from the very beginning of our program's execution. Useful mostly for people developing these parts of opsbox.
  • --verbose, which allows you to see the data that flows in certain parts of the application on the CLI.

By default, we output to stdout, and we specify the "INFO" logging level.

Getting Help

Opsbox has a colorful help system that can lend a helping hand.

To access general, application wide parameters and help, run opsbox --help without any modules specified.

If you'd like to see the arguments for a specific pipeline, run opsbox --help with the pipeline you want specified.

If you'd like to see all plugins from the help screen, add the --see_all argument.

Working with local plugins

When developing plugins, you will want to point to the directory in which they can be found.

This can be done using the --plugin_dir argument, which should point to a directory that contains a plugin in its or its children's subdirectories.

Buyer Beware

When specifying this argument, you will lose access to plugins stored in your virtual environment.

In addition, this argument requires that all plugin dependencies be installed in your executing virtual environment.