Tasker is a task runner, plain and simple. Define tasks, schedule them, and that's it !
Give it a try ! Just create the following docker-compose.yml
file
version: "2"
services:
tasker:
image: strm/tasker
volumes:
- "/var/run/docker.sock:/var/run/docker.sock"
environment:
configuration: |
schedule:
- every: minute
task: hello
tasks:
docker:
- name: hello
image: debian:jessie
script:
- echo Hello world from Tasker
And that's it, now you have a task, running inside docker every minute, using the debian:jessie
image (and bash inside this image), to run the script defined inside script
element.
Configuration is segmented, the concept is, you first configure your tasks, and then you configure what will activate them. Then just map your configuration file to /application.yml
in the container, or if you run it outside docker, map it to a file called application.yml
in the same folder that you are running the application from.
Here is an example with several constructions:
config:
global-environment:
- http_proxy="http://yourproxy:8080"
schedules:
- every: 5 minutes
task: test
- every: Saturday
task: backup
- cron: 0 0 13 * * *
task: test
tasks:
docker:
- name: backup
image: debian:jessie
script-strict: true
script:
- echo Running the backup
environment:
- TEST=environment variable value
volumes:
- someVolume:/whatWouldBeMappedInTheContainer
At this moment there is only one kind of task, docker
tasks. Those tasks ran inside docker containers, plain and simple as that. More task executers will be implemented in a near future, watch this repository to be sure you get the updates !
Configurations
-
image
- the image in the very same format as it is expressed for docker, inrepo/image:tag
for images in the default repository, ofserver/repo/image:tag
for images residing somewhere else. -
environment
- Define environment varibales to be used in the task execution, they follow the same pattern that you use indocker-compose.yml
file, a list ofvariable=value
. -
volumes
- An array, just like you mapvolumes
in yourdocker-compose.yml
. -
ports
- An array, just like you mapports
in yourdocker-compose.yml
. -
network
- If you wish to attach or use any other network that you have. If the desired network doesn't exist, it will be created.
Configurations regarding container life cicle:
keepContainerAfterExecution
- Keep container after it executes, won't delete it, WARNING it can leave a lot of trash. USE IT FOR DEBUGING ONLY !always-pull
- A boolean (true/false) property, when it's true, Tasker will pull a newer image version updating it if there is a newer one available.reuse-container
- If the container doesn't exist, create it, if there is already a container with that name, reuse the container.
You can pass parameters to your task and set the entrypoint of the image as you pass in docker command line. Example:
tasks:
docker:
- name: hello
image: debian:jessie
entrypoint: /bin/bash
arguments: -c,echo Aloha world
The example above execute /bin/bash
passing as argument -c echo Aloha world
. Comma can be used to separate arguments in an inline array. It uses a YML structure, so you can rewrite it as you wish. For example, you can declare the parameter array as a list:
tasks:
docker:
- name: hello
image: debian:jessie
entrypoint: /bin/bash
arguments:
- -c
- echo green bar
Both examples will produce the very same result.
script
directive is a facilitator to setting the entrypoint
to /bin/sh
and pass as arguments as a list of commands. It's syntax is just a list of commands like:
- name: helloScriptStrict
image: debian:jessie
script:
- echo Hello from Docker
Commands will be executed sequentially, no matter the result. To enable the strict mode, where the next command will only be executed if the previous command was successful (exit status 0), you can use the script-strict
property. For example
- name: helloScriptStrict
image: debian:jessie
script-strict: true
script:
- echo This is the first line
- echo This second line will only be executed if the above command properly runs
Tasker uses the concept of lazy loading, in other words, it will pull your image when it will run at the very first time. You can use always-pull=true
in your docker task definition if you want to keep updating the local image everytime your task runs.
- Tasker can run in a swarm or single instance, but it will schedule tasks to run in the current node.
- You can set
DOCKER_HOST
environment variable to this image, to make it run the scheduled tasks in a remote docker host. secrets
aren't available at the moment
Scheduler configuration is an important aspect in Tasker. The scheduler is responsible to what it's name suggests, schedule tasks to be executed. It is defined in schedule
section of the configuration file. Example
schedule:
- every: 10 minutes
task: test
- cron: 00 00 11 * * *
task: test
schedule
is an array of schedules. There are two ways to schedule an task. One is by giving it an interval using every
property or cron
. You can name a schedule with the name
attribute, but it isn't mandatory, you can have anonymous schedules if you prefer.
Every is a representation for a more simple understanding, like every: minute
, is far more readable, but won't give you the same power as the cron
directive. The every
parameters can be:
- Minute - Every minute inside one hour period. For example,
every: 10 minutes
, will result in 6 executions, at 00, 10, 20, 30, 40 and 50. Every minute scheduled task will start it's first execution at the first minute of the hour (00). - Hour - Every hour inside one day period. For example,
every: 6 hours
will result in 4 executions, at 00, 06, 12 and 18. Every hour scheduled task will start it's execution at the very first hour of the day (00 - midnight). - Weekday - Every weekday as in the list (Monday, Tuesday, Wednesday, Thursday, Friday, Saturday and Sunday). The task will be executed at the first minute of the first hour of the respective day. For example
every: monday
. - Year - Will be executed in the very first second, of the first minute, of the first day, of the first month of the year.
cron
is another way to set the expected trigger to your scheduled task. It obey the standard unix
cron format, but with the exception that it has one more field. Is a field representing seconds
, and is the first field, from 00
to 59
are the accepted values.
Is possible to configure Tasker to notify other components when tasks finish. Notifications are triggered after the task is completed. Notifications are defined in the notify
section of the configuration file, separated for each sub-type of notification, for example:
notify:
email:
- name: notifyEmailTest
task: helloNotifyEmail
when: always
server: mail.gmail.com
subject: Email subject
sender: [email protected]
content: Your message
Notifications need to be linked to a task, so a valid task
should be informed.
name
- A name to identify the notification.task
- Which task will trigger this notificationwhen
- When the trigger will run. There are three scenarios,always
- Is the default behavior, it will trigger the notification every time that the task runs.on-success
- Will trigger the notification only if the task is successful.on-error
- Will trigger the notification only if the task isn't successful.
In Tasker, e-mail notifications are a sub-category of notifications. E-mail notifications are configured with email:
field, inside the notify
section. Bellow a simple example configuration using Gmail:
notify:
email:
- name: notifyEmailTest
task: helloNotifyEmail
server: smtp.gmail.com
username: [email protected]
password: yourpassword
recipients: [email protected]
subject: Email subject
content: Your message
A complete example bellow
notify:
email:
- name: notifyMyTaskOnErrorOnly
task: yourTaskName
when: on-error
server: localhost
port: 3025
username: [email protected]
password: yourUserPassword
protocol: smtp
starttls: true
debug: true
subject: Email subject
recipients:
- [email protected]
- [email protected]
content: |
Multiline
Task finished
Configuration parameters regarding the connection to the e-mail server:
server
- Remote server where the email server is running. Defaultlocalhost
.port
- Remote port where the server is listening to. Default587
.username
- Username that will be used in the authentication with the server.password
- Password for the given username.protocol
- Protocol to use for e-mail transfer. Defaultsmtp
.starttls
- Use TLS or not. Default istrue
.debug
- Enable a more verbose output of the e-mail transfer process, use it only for troubleshooting. Defaultfalse
.validate-server-on-configuration-load
- Validate the e-mail configuration when Tasker startup, and abort it if isn't possible to connect to the e-mail server. Defaulttrue
.
Configuration parameters regarding the e-mail to be sent:
subject
- The subject of the e-mail.sender
- TheFROM
field in the e-mail.recipients
- A list of e-mail addresses that will receive the e-mail.content
- The content of the e-mail.template
- A velocity template that will be rendered and will generate thecontent
of the message.
If you wish to use the template
property, you need to follow Velocity Template Language specification.
There are some variables that can be accessed in the context of the execution of the template, and those are:
success
- (Boolean) Will betrue
if the task was successful.error
- (Boolean) Will betrue
if the task wasn't successful.log
- (String) The output of the execution of the task.task
- (String) The name of the executed task.start
- (Date) The Java Date when the task started.end
- (Date) The Java Date when the task finished.
If you want to suppress some of the logging information, you can add the following to your configuration
logging:
level:
ROOT: WARN
org.springframework.web: WARN
Configuration is the place where additional components of Tasker can be configured.
global-environment
- A list of environment variables that will be assigned to all tasks. Note that this variables are overwritten byenvironment
variables section of each task in case of a conflict.