You are viewing a plain text version of this content. The canonical link for it is here.
Posted to notifications@apisix.apache.org by "luarx (via GitHub)" <gi...@apache.org> on 2026/06/15 12:54:26 UTC

[I] Apisix-ingress-controller chart missing HPA template — autoscaling values have no effect [apisix-helm-chart]

luarx opened a new issue, #977:
URL: https://github.com/apache/apisix-helm-chart/issues/977

   Chart version: apisix-ingress-controller v1.2.0
   Controller version: 2.1.0
   
   The chart's values.yaml documents an autoscaling block:
   
   ```
   autoscaling:
       enabled: true
       minReplicas: 2
       maxReplicas: 3
       targetCPUUtilizationPercentage: 80
       targetMemoryUtilizationPercentage: 80
   ```
   
   However, there is no corresponding hpa.yaml template in charts/apisix-ingress-controller/templates/. As a result, enabling autoscaling has no effect — no HorizontalPodAutoscaler is created and the deployment stays at deployment.replicas (default 1).
   
   The apisix chart (same repo) does include templates/hpa.yaml and renders an HPA correctly. Could the same template be added for the ingress controller chart?


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: notifications-unsubscribe@apisix.apache.org.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


Re: [I] Apisix-ingress-controller chart missing HPA template — autoscaling values have no effect [apisix-helm-chart]

Posted by "geographybuff (via GitHub) via notifications" <no...@apisix.apache.org>.
geographybuff commented on issue #977:
URL: https://github.com/apache/apisix-helm-chart/issues/977#issuecomment-6693006593

   Agreed that the template belongs here, and it looks like a regression rather than a gap: this chart had an `hpa.yaml` until #837 removed it in the v2 update, which left the `autoscaling` values and the `pdb.yaml` branch behind. #1031 restores it.
   
   One difference from the `apisix` chart you pointed to: this controller uses leader election, so only the leader reconciles. In a k3d test with 3 replicas, only the leader started workers, and a standby took over about 30 s after the leader pod was deleted. So an HPA here gives failover rather than more capacity. A small range like your 2 to 3 fits that, and the PR defaults `maxReplicas` to 3 instead of the `apisix` chart's 100. Whether to offer an HPA on that basis is a maintainer call, and the PR lists removing the values as the alternative.
   
   AI disclosure: written with Claude Code.
   


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: notifications-unsubscribe@apisix.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org