Exclude terminating and Failed machines/nodes from DWD probe calculation of Failed Leases

[ashwani2k]

How to categorize this issue?

/area disaster-recovery /area robustness /kind enhancement /priority 1

What would you like to be added: Currently in DWD failed lease calculation all the nodes are considered to arrive at the nodeLeaseFailureFraction. This can be misleading for cases like:

1. When we have nodes which are in Terminating or Failed phase esp. if they will take the entire machineDrainTimeout due to PDB or other issues with eviction of pods.
2. Machines which are in crashLoopBackOff as they may never get created and including them in the count might also not be correct.

Why is this needed: To avoid DWD to mistakenly initiate meltdown protection for clusters which are quick to hit the nodeLeaseFailureFraction if they are having prolonged occurrence of the above 2 phases for nodes/machines. Also observed in issue-live(4796)

[gardener-ci-robot]

The Gardener project currently lacks enough active contributors to adequately respond to all issues. This bot triages issues according to the following rules:

• After 90d of inactivity, lifecycle/stale is applied
• After 30d of inactivity since lifecycle/stale was applied, lifecycle/rotten is applied
• After 30d of inactivity since lifecycle/rotten was applied, the issue is closed

You can:

• Mark this issue as fresh with /remove-lifecycle stale
• Mark this issue as rotten with /lifecycle rotten
• Close this issue with /close

/lifecycle stale

[unmarshall]

/remove-lifecycle stale