Rudder

Jonathan CLARKE wrote:

I agree with this. I can tell you that the code in Policy Templates to manage this is mostly ready to manage this on a node by node basis, so I don't think we should remove. I don't know about the code in the webapp though, I'll leave that up to you.

For now, there is no way in the webapp to mark a group or a node as "not managed".
To be consistent with the real node state regarding CFEngine, we need to use the same logic to check if a node is currently in "not managed" mode or not. I believe that that state is marked thanks to a given file (promise) in some directory, and so we can't use any sort of memory/db cache of it (the "red button" feature could be use from the command line or the file put by hand for a given node)

Once that is done, we can start to think to a group-level "red button", whose actual state would be the state of each of its node. The problem is that checking that state may be quite inefficient, as it would imply to check for perhaps hundred of file presence.

So, in the end, I do think that not much of the actual logic in the webapp can be keep, and that their is quite some specification and implementation work remaining, even for the node-only "red button", or to find a 80%-good solution.

The CFEngine code is indeed valid (and on a node basis)
As for the webapp part, i'm afraid we are a bit late in the release cycle to safely remove it; even more since there are event related to it, and scripts run by it.

It is still there is the system techniques, displaying a pretty non-understandable:

E| compliant     Common            Red Button               Red Button is not in effect, continuing as normal...

at every run on every node.

Done in 5.1 (#14045)