KCP: expose additional LeaderElection configuration as flags #2653
Labels
area/control-plane
Issues or PRs related to control-plane lifecycle management
kind/feature
Categorizes issue or PR as related to a new feature.
lifecycle/active
Indicates that an issue or PR is actively being worked on by a contributor.
Background
The default leader election configuration options are:
LeaseDuration
: 15 secondsRenewDeadline
: 10 secondsRetryPeriod
: 2 secondsThese values are quite small, but reasonable for most controllers. However, when the KCP controller is facilitating an upgrade, there is a legitimate risk that API server connectivity is disrupted (see #2652). On a self-managed cluster, because of these small values, this disruption can result in the controller losing track of its leader election. This causes the controller pod to crash.
The Request
We (as self-managed cluster users) would like to be able to configure these parameters. Flags seem reasonable!
/kind feature
/cc @vincepri @mytunguyen @sethp-nr
The text was updated successfully, but these errors were encountered: