HPA configuration error

Hi team!

We are trying to configure the HPA (Horizontal Pod Autoscaler) for a DB-less Kong Ingress Controller with a Helm upgrade with the following values.yaml:

autoscaling:
enabled: true
minReplicas: 3
maxReplicas: 5
targetCPUUtilizationPercentage: 80

Our:

autoscaling:
enabled: true
minReplicas: 3
maxReplicas: 5

resources:
limits:
cpu: 2
memory: 3G
requests:
cpu: 500m
memory: 300Mi

and with both configuration in the kubernetes cluster the HPA remains in the following state:

NAME REFERENCE TARGETS MINPODS MAXPODS REPLICAS AGE
kong-kong Deployment/kong-kong /80% 3 5 3 13d

and I get the following warning message:

failed to get cpu utilization: missing request for cpu

Here some complementary information:

-Kubernetes cluster: AKS
-Helm chart version: kong-1.15.1
-App Version: 2.3
-Helm config reference file: charts/values.yaml at main · Kong/charts · GitHub
thanks in advance for your help.

Regards,

CZ

Resource configuration is per-container, so you’ll need to make sure that you’ve set limits for both the Kong container and controller container. The controller is pretty lightweight and doesn’t require high resource requests.

thank you very much! that solved the issue. :grinning:


© 2019 Kong Inc.    Terms  •  Privacy  •  FAQ