Create second pipeline configuration question

We hired a dev to come build our SG integrated pipeline. It consists of 3 tools launched from SG Desktop. That dev is no longer available to us. I have since then added multiple api tools that interact with SG but do not use toolkit. They are just standalone python tools that do a lot of work on the farm and update SG using the api. I’d love to learn more about using SG Desktop and pipeline configurations but am terrified of doing something that interrupts our production. I feel like if I did something wrong I wouldn’t know how to very quickly undo something. We are in peak production and can not have any downtime. Is it safe to create other pipeline configurations? Is there a way to make sure that whatever I do does not become ‘live’ to productions? I feel like I don’t even know the proper terminology to ask.

So, I’ll just ask. If I screw up creating a new pipeline configuration will it impact our production?


Yes, you can use a separate pipeline configuration for development, with the config residing e.g. on a local drive.
Initialization and startup — tk-core v0.20.1 documentation explains how to do this. The app you are developing can also be on your local drive (if you wish) and will only get pushed to the central repository once you are satisfied with it. This workflow is really compulsory if you are doing anything but the simplest changes.

Also have a look at Developing apps - SG Developer - wrapping your logic in toolkit apps gives you several things for free, such as access to toolkit and shotgun api instances. I was also anxious before trying it, and it was quite beneficial.

All that said, there are still risks - the development might require changes to templates and env configurations, which when deployed to production might have unforeseen consequences. Testing thoroughly can mitigate some of this.