Comments (1)
I think it would be great if we could reuse https://github.com/aws/aws-node-termination-handler, no need in reinventing the wheel here.
As far as how it would integrate is something up for debate.
- NTH already handles events via EventBridge to SQS, so we could add a karpenter event that causes NTH to drain a node. NTH could even be configured to send a webhook to karpenter when it's complete for the node deletion. This would be the easiest, since all the functionality exists to do this today.
- Functionality could be added to NTH to make it more like a proper k8s controller. Karpenter could then mark nodes as needing to be drained, NTH would see nodes tainted or labeled a certain way (or monitor the Node CRDs), drain them, and then remove the label - something like that... This seems more extensible for other termination handler components to be plugged in later if necessary.
- Karpenter could call NTH directly via some registered http handler... this option doesn't seem great, but could work.
from karpenter-provider-aws.
Related Issues (20)
- conflict in policy between two docs HOT 4
- Delay scale-out of provisioning of Nodes by queing number of "Pending" Pods without batching
- What's karpenter EC2 selection policy if the workloads overcommit heavily (limits/request>3) ? HOT 1
- Can not properly delete NodeClass HOT 1
- Getting Started with Karpenter HOT 1
- Karpenter drift not working after update HOT 1
- Can't deploy Karpenter 0.37.0 to EKS 1.30 via EKS blueprints HOT 1
- Scheduling issue for g5 instance HOT 1
- Karpenter Fails to Fall Back to On-Demand in Local Zone with InvalidFleetConfiguration Error HOT 2
- Consider SaaS node-based pricing in Karpenter scheduler for cost optimization HOT 1
- custom label that does not use a restricted domain HOT 2
- Change to image tagging stratgey HOT 2
- Migrating pods from One NodePool to Another HOT 1
- `make setup` has been deleted but still exists in the document
- GPU workload pod failed to schedule due to low-priority gpu placeholder daemonset HOT 2
- Allocatable memory less than actual memory during scheduling (trn1.32x) HOT 1
- Upgrade process for Karpenter HOT 1
- Support pod PID limits through NodePool spec.template.spec.kubelet
- EC2 capacity blocks for ML in Karpenter Nodepools
- Pods were CrashLoopBackOff in provisioned node by karpenter(rare occurrences)
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from karpenter-provider-aws.