Node scaling failing / problems

[Yann (Yann T)]

Hi

Kube version: 1.21

I’ve a cluster scaling problem.
My backend app is one app: svc-platform, a one postgres 14 db: qovery-platform-preview-db.
I’ve a “review-template” env with those, and I’ve enabled auto preview-env on that preview-template.

I have a 0.5 CPU limit on my svc-platform, and I use t3medium nodes (2CPU, approx 1.6 usable I guess). I know I have a CPU spike at startup of my app, probably due to running database migrations, but then the app is barely using CPU. I know though that kube will provision the 0.5 cpu I’ve set as a limit though

Since yesterday all my deployment of svc-platform are failing due to:

I’ve then raised the limit of nodes in my cluster from 3-5 to 3-10:

but all my deployments continue to fail, and I have observed a maximum of 5 nodes in my cluster, not scaling above:

Can you help me understand why my cluster is not scaling above 5 nodes, and make all my deployment fails ?

[bchastanier (Benjamin Chastanier)]

Hey @Yann,

Sorry for the delay here, let me have a look tomorrow

Can you give me the link to your console so I can dig?

Did you triggered cluster “update” after applying settings changes on the cluster?

Cheers

[Yann (Yann T)]

Hi @bchastanier, I must have not trigger the update before, I thought so but I retried this afternoon and this time the cluster is matching my new node spec (I forced it to 6 node min to be sure to see a difference, and have indeed those 6 nodes).

I went back and forth between V2 and V3 in several chrome tabs, maybe it had some effect ? (in the V3 my cluster option UI seems not to work)

I’ll get it back down to fewer min nodes, and see if I get auto-scaling problems then