Are there any plans to automate node management in the future?
Currently, when governance by the NNS stops, the number of nodes can only decrease, and the ICP network will eventually shut down.
Welcome to the community @nokemono !
I’m not sure what you mean by this. Can you please elaborate? Based on which facts do you conclude this?
I may be misunderstanding, but I believe that the network will eventually shut down when the NNS stops because the NNS is managing things like adding and removing nodes.
I see. I think I understand the question better now, thank you. So the scenario you are thinking about is:
- the NNS governance will stop at some point because of a hardware/software problem?
- then… no one will pay attention to this, which will bring us to:
- regular subnet nodes will stop as well at some point because of a hardware/software problem
However, we do actually have a subnet recovery process in place (executed many times although luckily not for the NNS subnet). We also have a DRE team that is actively monitoring and managing the health of the IC nodes and submitting proposals to replace the unhealthy ones as needed. This way we always have a decent pool of ready-to-use nodes (not yet assigned to a subnet).
Did I get the concern correctly, and does the answer help?
3 Likes
Also, related to the title of this forum thread… Yes the plan is actually to move as much of the DRE activities into canisters and make the ICP network as self-managing as possible.
For this we recently added trustworthy node metrics, that will be the cornerstone of the future automatic unhealthy node replacements in subnets (you first need to know if a node is healthy or not, if you want to replace the unhealthy nodes).
So in summary… yes, you are right, this is high priority for the DRE team and we are actively working on making this happen.
5 Likes
