We scale up and down our self-hosted runners based on activity and schedule. When the runner host
(AWS EC2) is shutdown it’s terminated. When new hosts launch, they register their runners on startup.
This used to create a lot of runners obviously, so we added a regular task to delete old offline runners using the API. If we didn’t do this, we found new runners would fail to register when the total number of runners got to around 50 or so.
Now with the new authentication token method, we have a runner created in the GUI and then register each system/machine with a runner on startup, and that machine gets it’s own system id. But we cannot see the list of machines/system ids registered either via the UI or API. Over time the number of machines registered for each runner will continue to grow as we register new ones every morning.
Will this list be culled automatically? What will happen if this hidden list of machines continues to grow? Is there a way to clean out old machine records for machines that are offline? I can’t find anything about this in the docs and I’m worried that the list of growing machines will eventually cause new machines to stop registering, but this time there’s no way to clear them out.