Skip to content

Conversation

@bianchi2
Copy link
Contributor

@bianchi2 bianchi2 commented Oct 30, 2025

Problem

It is possible that escalator hits terminating nodes will breach minimum node size condition which prevents it from scaling down/deleting nodes in the nodegroup. This results in empty nodes running for no reason until ASG_DESIRED - NODES_TO_TERMINATE >= ASG_MIN (ie some manual intervention may be required)

Proposed Solution

While this situation can be caused mostly by bad timing and specific escalator configuration (increasing the time until a node would get deleted), it would be good to do some maths before tainting nodes, namely checking if ASG_DESIRED - NODES_TO_DELTE < ASG_MIN in other words run the same check as in DeleteNodes function but before nodes are tainted/marked for removal.

@bianchi2 bianchi2 changed the title Avoid tainting too many nodes Avoid tainting too many nodes if their deletion will breach ASG min Oct 30, 2025
@bianchi2 bianchi2 changed the title Avoid tainting too many nodes if their deletion will breach ASG min Avoid tainting too many nodes if their deletion breaches ASG min Oct 30, 2025
Copy link
Member

@vincentportella vincentportella left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

lgtm

@bianchi2 bianchi2 merged commit 80740ea into master Oct 31, 2025
6 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

5 participants