Thanks to visit codestin.com
Credit goes to github.com

Skip to content

configuration.md

Latest commit

 

History

History
374 lines (277 loc) · 8.33 KB

configuration.md

File metadata and controls

374 lines (277 loc) · 8.33 KB

Configuration format

The autoapply configuration is specified in YAML and has the following structure:

init:
  cwd: ...
  commands:
    - ...
loop:
  sleep: 60
  onerror: continue
  cwd: ...
  commands:
    - ...
server:
  enabled: false
  port: 3000
call:
  path: /path
  headers:
    - ...
  authentication:
    - ...
  stream: false
  onerror: fail
  cwd: ...
  commands:
    - ...

All sections are optional, but you have to configure at least one loop or call.

init

Define any commands that should be executed on startup. If any of the commands fail, autoapply will exit.

cwd

Directory in which to execute the commands (working directory). If unset, the directory in which autoapply was started will be used.

Example: cwd: /tmp

commands

List of commands that should be run on startup. See commands.

loop

Define the loop commands. If you only need one loop, you can just specify an object:

loop:
  cwd: /tmp
  commands:
    - ls

If you want to run multiple loops in parallel, you can specify an array:

loop:
  - cwd: .
    commands:
      - echo "This is loop 1"
  - sleep: 120
    commands:
      - echo "This is loop 2"

cwd

Directory in which to execute the commands (working directory). If unset, a new temporary directory will be created for each loop iteration

sleep

Time to wait (in seconds) between each iteration. A value of 0 disables sleeping, default is 60 (1 minute).

onerror

What should happen when a command fails? Possible values:

  • ignore - ignore the error and continue with the next command
  • continue (default) - don't execute any remaining commands and continue with the next loop iteration
  • fail - exit the loop with an error message

Example:

loop:
  onerror: ignore
  commands:
    - git pull
    - make &> .build
    - slackcli -h '#build' -m "Build status $(cat .build)"

commands

List of commands that should be run in a loop, after the commands defined in the init section have finished successfully. See commands.

server

Configuration of the internal HTTP server.

When the server is running, it provides the endpoint /healthz, which just responds with OK and can be used as a liveness probe.

enabled

Set this to true if you want to enable the HTTP server (default is false).

The server will be enabled automatically when you specify calls.

port

The HTTP port that the server will listen on, default is 3000.

call

Define web commands that will be executed when the URL is called. If you only need one call, you can just specify an object:

call:
  path: /echo
  commands: ["echo hello"]

If you want to provide multiple calls, you can specify an array:

call:
  - path: /echo1
    commands: ["echo hello"]
  - path: /echo2
    commands: ["echo world"]

The following environment variables will be set for each call:

  • REQUEST_METHOD - the HTTP method, for example GET or POST
  • REQUEST_URI - the requested URI, for example /echo1
  • REMOTE_ADDR - the IP address of the requesting client

Additionally, all request HTTP headers and query parameters will be available as environment variables.

Header name environment variables will be prefixed with HTTP_ and query parameters will be prefixed with QUERY_. Both header names and query parameters will be converted to uppercase and all characters other than letters or numbers will be replaced with _.

For example, the HTTP header User-Agent will be available as HTTP_USER_AGENT, the query parameter from the URL /test?param.1=x will be available as QUERY_PARAM_1.

path

The path to bind this call to:

call:
  - path: /date
    commands:
      - date

headers

HTTP headers to send with each response

call:
  - path: /index.html
    headers:
      - name: Content-Type
        value: text/plain; charset=utf-8
    commands:
      - echo '<html><body>Hello!</body></html>'

You could also pass the headers as an object:

headers:
  Content-Type: text/plain; charset=utf-8

methods

List of HTTP methods that this call accepts. By default, only GET methods are supported, requests with any other method will be answered by HTTP error 405 (method not allowed).

To support GET and POST:

call:
  - path: /clear
    methods: ["GET", "POST"]
    commands:
      - rm -rf /tmp/*

If you want to support all methods:

call:
  - path: /check
    methods: ["*"]
    commands:
      - echo $REQUEST_METHOD

authentication

List of users that are allowed to access this call. The users will be authenticated using HTTP basic access authentication and when invalid credentials are given, the call will be answered with HTTP error 401 (unauthorized).

call:
  - path: /private
    authentication:
      - username: user1
        password: password123
      - username: user2
        password: password321
    commands:
      - cat /var/log/access.log

The list of users can also be read from an environment variable:

call:
  - path: /private
    authentication:
      env: USER_AUTH
    commands:
      - cat /var/log/access.log

It's also possible to read the users from a file:

call:
  - path: /private
    authentication:
      file: /etc/users.txt
    commands:
      - cat /var/log/access.log

For environment variables and files, use the following format:

user1:password123
user2:password321

stream

Start to send available data to the client before all commands have finished? Default is false.

Data is sent before the commands have finished, so when using stream, the HTTP status will always be 200 (OK), regardless of any errors.

call:
  - path: /access.log
    stream: true
    headers:
      # See https://stackoverflow.com/a/35848615
      - name: X-Content-Type-Options
        value: nosniff
    commands:
      - tail -f /var/log/access.log

cwd

Directory in which to execute the commands (working directory). If unset, a new temporary directory will be created for each call.

onerror

What should happen when a command fails? Possible values:

  • ignore - ignore the error and continue with the next command
  • continue - don't execute any remaining commands and return HTTP status 200 (OK)
  • fail (default) - stop the call and return HTTP status 500 (error)

Example:

call:
  - path: /check
    onerror: ignore
    commands:
      - curl https://example.com/test1
      - curl https://example.com/test2
      - curl https://example.com/test3

commands

List of commands that should be run when the URL is called. See commands.

Common

commands

An array of commands to run.

The simple string form will use the shell, so environment variables etc. can be used:

commands:
  - ls ${HOME}
  - "echo $(pwd)"

The array form will not use the shell. In this case, ls will try to list the contents of a directory with the literal name ${HOME}. This will likely fail with "No such file or directory":

commands:
  - ["ls", "${HOME}"]

It is also possible to specify full scripts to be executed:

commands:
  - script: |
      #!/usr/bin/env node
      console.log('Hello, world!');

stdout

How to treat the standard output of the command? Possible values:

  • pipe (default) - print the output as the standard output of the autoapply process
  • ignore - discard any output (like redirecting to /dev/null)

Example:

commands:
  - command: ls "${HOME}/.ssh/id_rsa"
    stdout: ignore

stderr

The same option as stdout, but for the error output of the command:

commands:
  - command: curl "$URL"
    stdout: ignore
    stderr: ignore