In the instance group definitions, we can specify a max and min size of instance. Why? In AWS, kops doesn't configure autoscaling to scale up/add new instances on load (for example) so having a min of 1 and max of 10 would constantly have a count of 1.
huang-jy
All 2 comments
I'm pretty sure that the min and max are there so you have the option of adding in auto-scaling. There are already many different projects trying to implement auto-scaling for kubernetes clusters, which works with the AWS Auto-scaling Groups. Kops also offers a cluster auto-scaler, though I haven't tried it out.
So if you don't add an auto-scaler then the max has no effect
FrederikNS
on 15 Jan 2018
Yes, I had to configure autoscaling on the ELB manually, after it was created. Is there an option in kops to configure it at the time of creation?
EDIT: https://github.com/kubernetes/kops/tree/master/addons/cluster-autoscaler
Looks like I hadn't installed this addon :man_facepalming:
huang-jy
on 15 Jan 2018
Related issues
thejsj
路
4Comments
endejoli
路
4Comments
justinsb
路
4Comments
olalonde
路
4Comments
mikejoh
路
3Comments
Most helpful comment
Yes, I had to configure autoscaling on the ELB manually, after it was created. Is there an option in kops to configure it at the time of creation?
EDIT: https://github.com/kubernetes/kops/tree/master/addons/cluster-autoscaler
Looks like I hadn't installed this addon :man_facepalming: