We are using Netflix Conductor for orchestration purpose. We have some customisation around conductor library.
We started to face issues after integrated with MySQL database. All Netflix instances sharing same database, when some instances start the workflow other instances also has visibility to that workflow execution.
Similar issue exists for conductor clients and that is handled via task domain
I know It’s very much useful for production when it comes to stage many stages running conductor and some of them has older build running. I have no control over which instance process the entire workflow. This causes issue when Sometimes we wanted to test conductor with our changes but we don’t have control over which instances processes the workflow.
For e.g. we made some changes on es6 indexes when we trigger workflow initially indexes created properly later it goes to some other instance and it’s creating another index with different pattern.
Let me know your suggestions to overcome this problem