sap / node-refiner Goto Github PK

A violation against the OSS Rules of Play has been detected.

Rule ID: rl-assigned_teams-3
Explanation: Does it have enough admins on GitHub? No

Find more information at: https://sap.github.io/fosstars-rating-core/oss_rules_of_play_rating.html

A violation against the OSS Rules of Play has been detected.

Rule ID: rl-assigned_teams-5
Explanation: Does teams have enough members on GitHub? No

Find more information at: https://sap.github.io/fosstars-rating-core/oss_rules_of_play_rating.html

Is your feature request related to a problem? Please describe.

The current way of handling the minimum number of nodes is by setting a minimum value in the config map that the node refiner operator reads. Sometimes the minimum required nodes are not very obvious, especially when Pod affinity and pod anti-affinity specifications. These specifications might lead to the side effect of having a minimum number of nodes set less than the needed minimum number of nodes the cluster needs to satisfy affinity specifications.

If the utilization is low, the node refiner will try to scale the cluster down and the cluster autoscaler to scale up to satisfy the specifications, leading to a sylo of scaling down and up.

Describe the solution you'd like

Automatically read the affinity and anti-affinity specifications and calculate the minimum number of nodes required by these specs.
Keep the higher number of minimum nodes from the results of the automatically calculated value vs. the pre-set number in the config map.

Describe alternatives you've considered

In practice and when trying the node refiner out, this wasn't an issue that came up, but it came to my attention while working on the project; thus, its presence is an improvement.

This could also mean that other reasons for sylo might come up after implementing this.