arpx 0.3.0

Automate and relate multiple processes.
Documentation

arpx runs PROCESSES as they are defined in PROFILES. PROFILES allow for the configuration of PROCESSES, MONITORS, and ACTIONS, which work together to execute commands, watch for triggering conditions, and respond to those triggering conditions with new commands.

PROFILES can be run entirely or a single PROCESS can be run from within a PROFILE. If no PROFILE is defined, arpx will look for a file named arpx.yaml within the current working directory to run. PROCESSES can be configured to execute alone (blocking), run concurrently, depend on other PROCESSES, and perform ACTIONS when specific conditions are met.

Command Info
-h, --help Output a usage message and exit.
-V Output the version number of arpx and exit.

Options

Command Info
-f PROFILE, --file=PROFILE Execute a PROFILE at the given filepath. Defaults to ./arpx.yaml.
-p PROCESS, --process=PROCESS Execute a single PROCESS from within the current PROFILE.

Profile configuration

PROFILES can be named arpx.yaml or formatted as <my-prefix>.arpx.yaml. PROFILES are currently the primary mode of configuration for arpx runtimes, at least until a more scriptable/less verbose interface is developed.

processes:                          // Define primary PROCESSES.
  - name: [NAME OF PROCESS]         // Add a unique name to identify the PROCESS within the arpx runtime.
    command: [COMMAND]              // The command to execute.
    cwd: [PATH]                     // Directory in which to execute command.
    blocking: [TRUE|(FALSE)]        // Whether the PROCESS should block the main thread or run concurrently.
    silent: [TRUE|(FALSE)]          // Whether to silence logs for the PROCESS.
    onsucceed: [ACTION]             // ACTION to execute if PROCESS exits with success.
    onfail: [ACTION]                // ACTION to execute if PROCESS exits with failure.

monitors:                           // Configure MONITORS for specific PROCESSES.
  - process: [NAME OF PROCESS]      // Specify the PROCESS to MONITOR.
    condition: [COMMAND]            // Define bash condition which, on successful exit, triggers ACTIONS.
    actions:                        // Specify ACTIONS to execute when triggering conditions are met.
      [ACTIONS]                     // See `Actions` section below for an overview on built-in and custom ACTIONS.

actions:                            // Define custom ACTIONS which can be activated by MONITORS under triggering conditions.
  - name: [NAME OF ACTION]          // Add a unique name to identify the ACTION within the arpx runtime.
    command: [COMMAND]              // The command to execute.
    cwd: [PATH]                     // Directory in which to execute command.
    silent: [TRUE|(FALSE)]          // Whether to silence logs for the ACTION.
    stdin: [STDIN]                  // Enter some stdin when the ACTION is triggered.
    onsucceed: [ACTION]             // ACTION to execute if ACTION exits with success.
    onfail: [ACTION]                // ACTION to execute if ACTION exits with failure.

Example profile - script.arpx.yaml:

processes:
  - name: loop1
    command: |
      for i in {1..5}
      do
        sleep 1
        echo "Loop1 $i"
      done
  - name: loop3
    command: |
      for i in {1..5}
      do
        sleep 1
        echo "Loop3 $i"
      done

monitors:
  - process: loop1
    condition: '[[ "$LOG_LINE" =~ "Loop1 5" ]]'
    actions:
      - loop2

actions:
  - name: loop2
    command: |
      for i in {1..3}
      do
        sleep 1
        echo "Loop2 $i"
      done
      exit

Example arpx concurrent output

processes:
  - name: loop1
    command: |
      for i in {1..5}
      do
        sleep 1
        echo "Loop1 $i"
      done
    blocking: true                  // Added
  - name: loop3
    command: |
      for i in {1..5}
      do
        sleep 1
        echo "Loop3 $i"
      done

monitors:
  - process: loop1
    condition: '[[ "$LOG_LINE" =~ "Loop1 5" ]]'
    actions:
      - loop2

actions:
  - name: loop2
    command: |
      for i in {1..3}
      do
        sleep 1
        echo "Loop2 $i"
      done
      exit

Example arpx blocking output

Processes

PROCESSES are the primary commands arpx will manage. PROCESSES can be run blockingly or concurrently, and can be run one at a time with the -p option.

To run an individual PROCESS named my-process contained in a file named arpx.yaml in the current working directory, execute:

$ arpx -p my-process

To run all PROCESSES contained in a file named my.arpx.yaml, execute:

$ arpx -f ~/path/to/my.arpx.yaml

Process properties:

  • name: Name of the PROCESS.
  • command: Command to execute when PROCESS is initiated.
  • cwd: Working directory in which to execute PROCESS command.
  • blocking: Whether the PROCESS should block the main thread or run concurrently. Blocking PROCESSES will run in the order in which they are defined in arpx.yaml, so long as they precede any non-blocking PROCESSES. Non-blocking PROCESSES will run concurrently to to other PROCESSES and will therefore not block the main thread from initiating the remainder of the defined PROCESSES. Setting blocking to true makes it possible to run PROCESSES in order, rather than concurrently.
  • silent: Whether the PROCESS should propagate stdout and stderr from the command to the arpx stdout.
  • onsucceed: An ACTION to initiate if the current PROCESS command exits successfully.
  • onfail: An ACTION to initiate if the current PROCESS command exits with failure.

Monitors

MONITORS watch for conditions in a given PROCESS and perform ACTIONS if/when those conditions are met. MONITORS are configured by defining a condition and actions to execute when the condition exits successfully.

Monitor properties:

  • process: Name of the PROCESS that this MONITOR's CONDITION should apply to.
  • condition: Shell condition which, upon a successful exit status, triggers ACTIONS. Conditions are checked with each line a program logs from stdout or stderr. The output of the current line is available within the condition script via the $LOG_LINE variable.
  • actions: List of ACTIONS to initiate if this MONITOR's CONDITION exits with success.

Actions

ACTIONS are new PROCESSES which can be executed when triggered by a MONITOR. There are built-in actions available to all MONITORS specified below. Custom ACTIONS can also be defined.

Built-in actions

  • exit: Exit arpx.
  • kill: Exit the current PROCESS. Does not work with onsucceed or onfail.
  • respawn: Exit and restart the current PROCESS. If used with onsucceed or onfail on an ACTION, arpx will attempt to spawn a PROCESS with a name matching the current ACTION.
  • silence: Silence the current log. Does not work with onsucceed or onfail.

Custom

Custom ACTIONS can define new tasks to be executed if/when triggering conditions are met. Currently, the only type of ACTION available is shell, which allows for defining a shell command to run when the current ACTION is activated. If ACTION is called via onsucceed or onfail, stdin will not apply.

Custom action properties:

  • name: Name of the ACTION.
  • command: Command to execute when ACTION is initiated.
  • cwd: Working directory in which to execute ACTION command.
  • silent: Whether the ACTION should propagate stdout and stderr from the command to the arpx stdout.
  • stdin: Enter some stdin. Allows for automatically replying to prompts from a PROCESS. stdin does not apply to ACTIONS called by onsucceed or onfail.
  • onsucceed: An ACTION to initiate if the current ACTION command exits successfully.
  • onfail: An ACTION to initiate if the current ACTION command exits with failure.

arpx is useful for automatically handling errors, suppressing stdout/stderr output, scheduling processes relative to each other, and more. It was originally concieved for the purpose of automatically handling errors while running a local development environment with multiple dependent services. The goal is that arpx should be generally useful, so please open issues for enhancements with a mind toward powerful, general utility.

Integrations

Programmatic

  • arpxjs: Programmatic process automation, relation, and multiplexing for Node.js