Microservices
Deployment
Microservice deployments are executed directly via Talend Remote Engine. For each microservice, the Data Service Runner starts a separate Java process as a subprocess of the Remote Engine itself.
Each microservice is deployed with a unique identifier and dynamically gets a unique HTTP port when started. It is displayed on the Task Details page in Talend Management Console after a successful deployment.
The deployment with a unique identifier and HTTP port allows you to deploy the same Talend Cloud artifact through multiple tasks with different configurations on the same Remote Engine.
- An artifact (the microservice ZIP based deployment)
- An application.properties file
- A context property file (depending on the Talend Cloud environment and the overwrite parameters set in Talend Management Console)
- Running a new Java sub-process of the current Talend Remote Engine
During deployment, each microservice is attributed a network port starting from 5060 by default.
The number of deployed microservices is limited to 20 by default. To change it, see Customizing global microservice configuration.
Undeployment
After a microservice is deployed, the only action linked to its unique identifier is undeployment. Undeployment means that every trace of its deployment is deleted from the <MicroserviceExecutionDirectory>. After undeployment, the network port is released and available for a new deployment.
Recovery
If Talend Remote Engine, the Data Service Runner, or the host machine is stopped, every deployed microservice is stopped and kept in the <MicroserviceExecutionDirectory>. When all services are started again, the microservice is started with the same configuration as before.
Any configuration update between a start/stop is ignored by the deployed microservices. To avoid recovery, microservices should be undeployed.