Once we have implemented and deployed an CloudConnect project to the GoodData platform, we'll take a look at the process execution, scheduling, and monitoring. Most of the use cases in this area are still implemented via the 'gray pages' that are less user convenient. The GoodData development team is currently working hard on an overarching CloudConnect project management console that enables convenient CloudConnect project execution, scheduling, and monitoring.
When an CloudConnect project is deployed to the GoodData platform, the process directory structure is archived and persisted under a target analytical project. The process is assigned with a unique URI during the deployment. The URI is used for identifying the CloudConnect project during it's execution, re-deployment, deletion etc.
Every time the CloudConnect project is executed, the process persisted directory structure is extracted to a temporary directory and executed. The temporary directory is located in a private Webdav store that is associated with the user who invoked the process. The user can access all intermediary files and logs that the process creates. The log files are persisted for 10 days after the CloudConnect project execution.
Because the CloudConnect project executions are isolated to the separate directories, the results of one process execution can't be automatically used in a subsequent execution of the same CloudConnect project. If the process needs to keep a state, it needs to use some explicit persistent store (e.g. EventStore)