Suraj Banakar(बानकर) | スラジ

ex-InfraCloud | ex-Quipper/Recruit | sig-autoscaling/OSS contributor | GSoC 2022 mentor | JLPT N1 | CKA/CKAD | likes motorcycles and nice people

[Suraj Banakar(बानकर) | スラジ @_vadasambar]

@aquaclivm Congratulations 🎉

[Suraj Banakar(बानकर) | スラジ @_vadasambar]

@sukukyon おめでとう🎉

[Shunsuke Suzuki @szkdash]

同僚じゃなくなったあとも OSS などソフトウェアを通じて繋がってられるのはいい話
ソフトウェアエンジニアの良いところでは

[Suraj Banakar(बानकर) | スラジ @_vadasambar]

TIL you can expose custom resources as kube state metrics: https://t.co/LhGSOn5xQU

[Suraj Banakar(बानकर) | スラジ @_vadasambar]

Kubernetes cluster-autoscaler (CA) now supports kwok cloudprovider which allows you to play around with (or test) the CA on your local machine without actually creating VMs in any cloud provider. One of my biggest PRs to CA. For details, check https://t.co/I3eHFwJTIc

[Suraj Banakar(बानकर) | スラジ @_vadasambar]

@wzshiming More info: https://t.co/HQXKgI0Sd2

[Suraj Banakar(बानकर) | スラジ @_vadasambar]

https://t.co/IPL9Mhsc0f made by @wzshiming scales up workloads from 0 when they get a request

[Suraj Banakar(बानकर) | スラジ @_vadasambar]

Is there a way to run cluster-autoscaler on your local `kind` cluster and make it scale your `kind` nodes? There is! https://t.co/hLm9N1ybXt

[Suraj Banakar(बानकर) | スラジ @_vadasambar]

If you are using cluster-autoscaler you might be interested in https://t.co/5FVAXpCxNR (looking for reviews/comments from users/contributors)

[Suraj Banakar(बानकर) | スラジ @_vadasambar]

PSA: Ran into an issue while trying to provision v1.26.1 `kind` cluster using cluster-api. Turns out `kind` does not have kindest/node with the tag v1.26.1 https://t.co/O0tHQwm1Y9
Related:
* https://t.co/SUgTWq5c4s

[Suraj Banakar(बानकर) | スラジ @_vadasambar]

Karpenter applied to be a part of sig-autoscaling yesterday:
* https://t.co/oAjuSh489L
* https://t.co/xXIjdOajMi

[Suraj Banakar(बानकर) | スラジ @_vadasambar]

@sanpo_shiho @kerthcet Congratulations 🎉

[Suraj Banakar(बानकर) | スラジ @_vadasambar]

Not sure if this is already well known but I saw an issue where a PDB was completely blocked. Turns out this was because one of the resources it was targeting was a DaemonSet which doesn't implement scale subresource.
- https://t.co/Bvoqr07vKZ
- https://t.co/3Ed6zryZD2

[Suraj Banakar(बानकर) | スラジ @_vadasambar]

@sanpo_shiho Looks interesting.

[Suraj Banakar(बानकर) | スラジ @_vadasambar]

Not documented but if you are using a custom cluster-autoscaler in GKE, you can set labels on the upcoming nodes using `AUTOSCALER_ENV_VARS` like this https://t.co/mFynS9CfkZ

[Suraj Banakar(बानकर) | スラジ @_vadasambar]

@ganjaaaaa お誕生日おめでとうございます!

[Suraj Banakar(बानकर) | スラジ @_vadasambar]

📝 Recently came to know that kubernetes scheduler supports extenders to extend scheduler logic via calling external http endpoints. `
docs: https://t.co/89J4Xr7AzO
example: https://t.co/yoN68urOiR

[Suraj Banakar(बानकर) | スラジ @_vadasambar]

@szkdash Maybe something to feature gates in Kubernetes might be a good idea https://t.co/FIA7jKYGd9

[Suraj Banakar(बानकर) | スラジ @_vadasambar]

New version of CA(cluster-autoscaler) will support adding pod annotation to skip local storage volumes to fix CA blocking node scale down.
新cluster-autoscalerバージョンでPodにannotationを加えてNodeをscale downするときlocal storage volumeがskipできます。
- https://t.co/oxWdmasZON