Trigger your Rundeck or Jenkins jobs from Slack.
Example:
@corebot deploy user-service 1.0 to staging
> corebot:
> OK, I'm deploying user-service version 1.0 to staging.
> Status of job is: running
> Details: http://rundeck/jobs/abc/123
Why would you want this? Check out ChatOps.
- As a Slack admin, create a Slack bot user and obtain its access token - instructions
- As a Rundeck admin, generate a Rundeck API token - instructions
- Set environment variables
- Run!
The quickest way to get up and running is to use our free cloud-hosted version at https://www.remotebot.io/bot
If you'd like to run Corebot yourself as a Docker container, you can do the following:
docker run -d \
--env SLACK_AUTH_TOKEN="CHANGEME" \
--env SLACK_CHANNELS="corebot" \
--env RUNDECK_API_TOKEN="CHANGEME" \
--env RUNDECK_BASE_URL="http://rundeck:4440" \
-v /path/to/actions.yml:/opt/corebot/actions.yml \
outofcoffee/corebot
Note: the container does not require any inbound ports to be exposed.
Note: See the Environment variables section for the available configuration settings.
As a Slack admin, create a 'Classic App' in Slack: https://api.slack.com/apps?new_classic_app=1
Add a bot user by expanding 'Add features and functionality', and clicking on Bots:
https://api.slack.com/apps/<your app ID>/app-home
Add the required scopes in the 'OAuth & Permissions' section:
https://api.slack.com/apps/<your app ID>/oauth
The scopes are:
chat:write:bot
Don't forget to save changes after adding scopes.
Install your app to your workspace. This will generate the token you need. You'll want to copy the 'Bot User OAuth Access Token'. It should look like this:
xoxp-123456789012-123456789012-abcdef1234567890abcdef1234567890
Corebot's capabilities are determined by its driver. For example, a driver might allow you to interact with a CI/CD system.
A modern version (1.7+) of Jenkins is required - version 2.x or higher is preferred.
Here is the official Jenkins Docker image:
docker run -it \
-p 8080:8080 \
jenkins
Example usage:
@corebot deploy services v2 to staging
> corebot:
> OK, I'm deploying user-service version 1.0 to staging.
> Status of job is: running
> Details: http://rundeck/jobs/abc/123
Any Rundeck instance can be used as long as it supports API v14 or higher.
As an example, here is an unofficial Rundeck Docker image: https://hub.docker.com/r/jordan/rundeck/
docker run -it \
-p 4440:4440 \
-e SERVER_URL=http://localhost:4440 \
jordan/rundeck
Example usage:
@corebot deploy services v2 to staging
> corebot:
> OK, I'm deploying user-service version 1.0 to staging.
> Status of job is: running
> Details: http://rundeck/jobs/abc/123
The items driver acts like a lending library. It allows users to borrow, return, or evict users from items in the library.
An example use case is a group of people reserving a server/environment.
Example usage:
@corebot list
> corebot:
> dev1 - no one is borrowing
> dev2 - in use by @alice, @bob
> staging - no one is borrowing
@corebot borrow dev1 for bugfixing
> corebot:
> OK, you've borrowed dev1
@corebot return dev1
> corebot:
> OK, you've returned dev1
Configure the bot using the following environment variables.
SLACK_AUTH_TOKEN="CHANGEME"
SLACK_CHANNELS="corebot"
ACTION_CONFIG_FILE="/path/to/actions.yaml"
Note:
SLACK_CHANNELS
is a comma-separated list of channel names, such as"channelA,channelB"
. You can also use regular expressions to match channel names, such as"channel.*"
.
Note: the default path for
ACTION_CONFIG_FILE
is/opt/corebot/actions.yml
. When using corebot within a Docker container, it is typical to add your configuration file at this location, or bind-mount a file to this path.
RUNDECK_API_TOKEN="CHANGEME"
RUNDECK_BASE_URL="http://rundeck:4440"
Note: ensure that the API token you generate in Rundeck has the necessary permissions to trigger builds. For more information, consult the Rundeck ACL documentation.
JENKINS_BASE_URL="http://localhost:8080"
JENKINS_USERNAME="CHANGEME"
JENKINS_PASSWORD="CHANGEME"
JENKINS_API_TOKEN="CHANGEME"
Note: typically you will specify the username and password for accessing a Jenkins instance. The token approach is rarely used and can be omitted.
Advanced variables to tune behaviour and performance:
CACHE_EXPIRY="60"
The cache expiry controls the period of time, in seconds, corebot holds the action configuration in memory after reading it from file.
EXECUTION_STATUS_TIMEOUT="120"
EXECUTION_STATUS_POLL="10"
The execution status timeout controls the period of time, in seconds, corebot will poll a running job for status updates, after which it gives up. You can also control the polling interval, also in seconds.
SLACK_REPLY_IN_THREAD="true"
Posts replies from the bot to a thread starting from the trigger message. Default: false
.
SLACK_ALLOW_THREADED_TRIGGERS="true"
Allows child thread messages to be trigger messages. Implies SLACK_REPLY_IN_THREAD="true"
. Default: false
.
CHAT_GENERATOR_FILE="/path/to/file.yml"
The path to an external YAML file containing custom chat lines. See the default file, default-chat.yml
, for examples.
SYSTEM_CONFIG_FILE="/path/to/file.yml"
The path to an external YAML file containing system configuration. See the sample file, system.yml
, for examples.
Corebot has both built-in operations and external actions. Examples of built in operations are the lock/unlock actions. External actions are triggers for your Rundeck/Jenkins jobs, configured using a configuration file, typically called actions.yml
.
Note: the configuration file path is specified with the
ACTION_CONFIG_FILE
environment variable.
Example file:
version: '1'
actions:
services:
jobId: 9374f1c8-7b3f-4145-8556-6b55551fb60f
template: deploy services {version} to {environment}
File structure:
- All files must specify version ‘1’
- All actions must sit under a top level
actions
block - Each action must have a name (it’s
services
in this example) - Each action must have either:
- a Rundeck job ID (obtain this from Rundeck), or
- a Jenkins job name
- Each action must have a template - more details below
- Each action may optionally specify a list of tags
- Each action may optionally specify option configuration, such as:
- static values
- value transformers
- whether the option can be locked
Tip: Check out the
examples/config
directory for sample configuration files.
Actions should specify a driver. The available drivers are:
- jenkins - use this to allow users to trigger Jenkins jobs
- rundeck - use this to allow users to trigger Rundeck jobs
- items - use this to allow users to borrow/return a set of items (e.g. books, servers etc.)
Note: if none is specified, the driver is assumed to be
rundeck
.
Example:
version: '1'
actions:
services:
driver: jenkins
jobId: my-jenkins-job
template: deploy web {version} to {environment}
Important: Ensure that you set the environment variables corresponding to the driver(s) you use, such as the base URL, API key/username etc.
An action template provides the syntax for invoking the command.
Example:
deploy services
A template also allows you to specify job options as placeholders, such as:
deploy services {version} to {environment}
In this example both version and environment are captured from the command, such as:
@corebot deploy services 1.0 to UAT
This will result in the action being fired, passing the following options:
- version=1.0
- environment=UAT
You might want to pass an option value to a job every time, and not require the user to provide it. You can accomplish this using the value
property of an option within the options
action block:
version: '1'
actions:
services:
jobId: 9374f1c8-7b3f-4145-8556-6b55551fb60f
template: deploy services {version} to {environment}
options:
myOption:
value: someValue
myOtherOption:
value: someOtherValue
This will result in the action being fired, passing the following options:
- version=1.0
- environment=UAT
- myOption=someValue
- myOtherOption=someOtherValue
You might want to transform an option value provided by a user before it is passed to a job. You can accomplish this using the transformers
section of an option within the options
action block:
version: '1'
actions:
services:
jobId: 9374f1c8-7b3f-4145-8556-6b55551fb60f
template: deploy services {version} to {environment}
options:
version:
transformers:
- LOWERCASE
environment:
transformers:
- UPPERCASE
If the user typed this command:
@corebot deploy services V1.0 to uat
This will result in the action being fired, passing the following options:
- version=v1.0 (note: lowercased)
- environment=UAT (note: uppercased)
You might want to lock an option value, so that it cannot be passed to an action.
Example: If you have an option to specify the environment for a deployment, you might wish to lock deployments to the environment named 'production'.
You can accomplish this using the lockable
property of an option within the options
action block:
version: '1'
actions:
services:
jobId: 9374f1c8-7b3f-4145-8556-6b55551fb60f
template: deploy services {version} to {environment}
options:
environment:
lockable: true
If the user typed this command:
@corebot lock environment prod
This will result in a lock being placed on the 'environment' option, with the value 'prod'.
With the lock applied, this will fail:
@corebot deploy services 1.0 to prod
...but this will still succeed:
@corebot deploy services 1.0 to uat
You can of course unlock the option with:
@corebot unlock environment prod
Note: It's strongly advisable to apply a transformer to lockable options, to ensure the value 'prod' is considered equivalent to 'PROD'.
Sometimes actions can be run on multiple jobs. To do this, set the tags
block:
version: '1'
actions:
deploy-services:
jobId: 9374f1c8-7b3f-4145-8556-6b55551fb60f
template: deploy services {version} to {environment}
tags:
- services
restart-services:
jobId: e9d12eec-abff-4780-89cd-56a48b8c67be
template: restart services in {environment}
tags:
- services
Here, two actions are defined: deploy-services
and restart-services
, both tagged with services
. This means you can do things like:
@corebot lock services
…and both actions will be locked.
Tip: There is a special tag set on all actions, named 'all'. This means you can do things like
@corebot lock all
.
Sometimes the bots reaction is enough to see the status. To do this, set the showJobOutcome
option to false
. Default is true
.
Sometimes the output of the job is needed to be given back by the bot. To do this, set the showJobOutput
option to true
. Default is false
.
version: '1'
actions:
deploy-services:
jobId: 9374f1c8-7b3f-4145-8556-6b55551fb60f
template: deploy services {version} to {environment}
showJobOutput: true
showJobOutcome: false
You can choose which users are authorised to perform actions, using the security
block:
security:
users:
# alice uses the built-in 'admin' role
alice:
roles:
- admin
Important: if you do not specify a security configuration explicitly, the default will be used. The default settings permit all users to perform all actions.
There is a built in role, named admin
, which you can assign to users. You can also define your own roles, listing the permissions granted to users with that role.
You can assign roles on a per-username basis or, if you wish to assign certain roles to all users, use the special "*"
key, as shown in the example below:
security:
roles:
# a role that can only trigger jobs
deployer:
permissions:
- trigger
users:
# alice uses the built-in 'admin' role
alice:
roles:
- admin
# all users can trigger jobs
"*":
roles:
- deployer
By default, role definitions apply to all actions. If you wish to restrict the permissions granted by a role to certain actions only, add a tag to the action and also to the corresponding tags
array in the role:
security:
roles:
# a role that can only trigger jobs
deployer:
permissions:
- trigger
# this role only permits triggering of actions tagged with 'services'
tags:
- services
There are a number of built in actions, such as:
@corebot help
- show usage information.@corebot lock {action or tag name}
- lock action(s) to prevent them being triggered accidentally.@corebot unlock {action or tag name}
- unlock locked action(s).@corebot status {action or tag name}
- show status of action(s).@corebot lock {option name} {option value}
- lock an option with a given value.@corebot unlock {option name} {option value}
- unlock an option with a given value.@corebot status {option name} {option value}
- show status of an option with a given value.@corebot enable {action or tag name}
- set the Rundeck execution status for a job - Note: this requires the Rundeck ACL to permit the API user to set the execution status of a job.@corebot disable {action or tag name}
- set the Rundeck execution status for a job - Note: this requires the Rundeck ACL to permit the API user to set the execution status of a job.
You can set some default operations for all actions using the system configuration file.
The path to this file is specified using the SYSTEM_CONFIG_FILE
environment variable.
An example configuration follows:
# System configuration
---
version: '1'
system:
defaults:
driver: jenkins
showJobOutput: false
showJobOutcome: true
runAsTriggerUser: false
options:
myVar:
value: someDefaultValue
requestHeaders:
Cookie: "foo=bar"
The example above sets a number of default values, which can be overridden by your individual action configurations.
This example also specifies a map of HTTP headers to set on each API request to the external build/deployment systems. These headers will be sent to Jenkins and Rundeck when making any requests.
Corebot supports plugins, loaded dynamically at startup.
The following plugin categories are supported:
- front-ends (e.g. Slack, WebSocket)
- back-ends (e.g. Rundeck, Jenkins, Items)
- stores (e.g. MySQL, Redis)
Plugins are packaged and distributed as WAR files and include all required dependencies.
You can load plugins by using the generic-bot
distribution.
See
bots/generic/README.md
for information. See theexamples/plugins
directory for plugin configuration examples.
If you wish to build and run locally, you can run:
./gradlew installDist
docker-compose build
Once built, set the environment variables in docker-compose.yml
. See the Environment variables section.
Then run with:
docker-compose up
If you change anything, don't forget to rebuild before running again.
Dependencies can be published to the project Maven repository.
Note: Publishing to the repository requires appropriate AWS keys to be set in
gradle.properties
.
The Corebot libraries are published to our Maven repository. You can include them as dependencies in your project.
To use the dependencies in a project, add the repository:
repositories {
maven {
url 'https://s3-eu-west-1.amazonaws.com/gatehillsoftware-maven/snapshots/'
}
// jitpack required for Slack dependency
maven { url "https://jitpack.io" }
}
...then add a dependency:
compile "com.gatehill.corebot:core:0.7.0"
Note: update the version with the latest stable release.
For recent changes see the Changelog, or view the Roadmap.
- Pull requests are welcome.
- PRs should target the
develop
branch. - Please run
ktlint
against the code first ;-)
Pete Cornish (outofcoffee@gmail.com)