Skip to content

otto-de/edison-jobtrigger

Repository files navigation

Edison Microservice is a library to write production-ready Microservices on top of Spring Boot. It is actively used at otto.de.

Because Microservices should be loosly coupled, they often need to import data from other systems - at least in our architecture. To make such kind of background jobs as easy to implement as possible, and to run them on a regular base in clustered environments, we need a component that is triggering these jobs as needed: that's the purpose of

Edison JobTrigger

All Edison Microservices may announce JobDefinitions: specifications about jobs telling a JobTrigger (or other tools) how a Job is triggered, in what frequency this should happen, and so on.

If a JobTrigger server is configured, the Microservice will register itself at startup and, in the following, update this registration every few minutes. The JobTrigger is querying all supported JobDefinitions and will trigger the jobs of all registered Microservices according to their definition.

Edison JobTrigger supports multiple services from different groups (or Teams) in multiple environments (test, prelive, live...).

That's it.

Usage

gradlew build
java -jar build/libs/edison-jobtrigger-<version>.jar

Write an Edison Microservice (or simply use Example-Jobs as an example) including some jobs and job definitions.

Configure some additional properties in your service:

edison.servicediscovery.servers=http://localhost:8080/jobtrigger # where to find the JobTrigger
edison.servicediscovery.service=http://localhost:${server.port}/${server.context-path} # where to find the service
# Optionally override defaults:
# edison.servicediscovery.expire-after=15 # expire the registration of the service after N minutes.
# edison.servicediscovery.refresh-after=5 # reload JobDefinitions after N minutes
# edison.servicediscovery.environment=unknown # the environment of the service (live, develop, ...)
# edison.servicediscovery.group=default # the group of the services. In our case it's the name of a team

Run JobTrigger + your service(s). After some time (one minute), JobTrigger will find the JobDefinitions of your services and start the triggers accordingly.

Discovery

Because JobTrigger has no persistency, Services have to re-register every few minutes: this is the purpose of property edison.servicediscovery.expire-after.

The known services and their JobDefinitions can be seen on the Discovery page: Discovery

Triggers

Because services can be deployed at any time, the JobTrigger will load current JobDefinitions every few minutes. The interval is configured using edison.servicediscovery.refresh-after.

After a while, you can see the triggered jobs on the Trigger page: Triggers

Job Information

You can access the asynchronously running job by opening the messages: Job Messages

Have fun!