I’d like to know if it is possible to specify certain sets of properties (variables) for each deployment target. Example, I have a batch processing environment where I have a lot of (30-50) worker nodes, each having a set of 5 properties (values being different for each node). I need to deploy a component (service) into each of these nodes, but update the config with the parameters. What is the best way to handle this right now? Using variable substitution will be clumsy as I will end up with almost 150-250 variables.
Thanks for reaching out. I’m afraid that because of the granularity of your setup, there’s no way to reduce the amount of variables you’ll need to create.
What you can do is use Tenants to help a bit with the variables organization. Instead of having all the variables in Projects -> [your project] -> Variables, you could do something like this:
Create a tenant per worker node (I’m assuming Worker node == Target) and add it to your project.
Add each Tenant to its proper Target.
Create a Project Variable Template on your project. This will make the project require all its tenants to fill the variables from that template to be able to deploy to them.
This way you’ll be able to manage each tenant’s variables on its own page, instead of having to deal with a single screen with 150-250 variables. See attached screenshots too.
Like I said, its not going to reduce the amount of variables you’ll need to use, buts its going to make it easier to manage.
