Is there a way to define a collection of agents with specific facts (server types, perhaps) and during an execution of a process pick one from that group instead of all. The use case would be a group of servers that have a specific utility installed, but the process needs to just run on any one of those.
Hello, yosef.
If I understood correctly, what you want can be accomplished after the deployment is created.
You can choose, for each process, on which server you want to run it.
Best regards,
Ricardo Bernardino
Thanks for the reply. By 'You can choose...' what exactly do you mean?
Hello.
I meant that you can assign specific servers to a process on a deployment, before you trigger it.
I attached a screenshot showing what I meant.
Hi,
Release Automation has two agent groups that are used for artifact retrieval and approval this allows you to define agent groups for these two specific functions and at run time one of the agents in the group will be used. So i think you are looking to expand this concept to allow you to create groups for other functionality, is this correct? If so this is not currently available in the product but would be a good suggestion for an Idea. As a work around could you create an action as part of the initialisation to dynamically assign agents?
Regards
Keith
@Keith, this would certainly be very nice to have. In many cases 'Utility Servers' should actually behave like the agent groups where only one is required to perform the functionality but it doesn't really matter which one. So for HA a group could be defined and an 'available' or 'least busy' one selected from the group at run time. Super cool would be to add to every group type a hook to determine availability or utilization stats for enabling RA to select the best choice for the job.