What steps did you take and what happened:
- Apply the following manifest (example from weight documentation):
apiVersion: projectcontour.io/v1
kind: HTTPProxy
metadata:
name: unsetting-weights
namespace: default
spec:
virtualhost:
fqdn: weights.bar.com
routes:
- services:
- name: s1
port: 80
- name: s2
port: 80
- Edit weights on s1 and s2
- Reapply the manifest
- Check the resource to see that the weights you edited in were removed
What did you expect to happen:
The weights are not managed by the manifest, so they should remain the same after step 3.
Anything else you would like to add:
We've been using this argo rollouts plugin to manage weights on our httpproxies. It changes the weights during canary rollouts gradually. Recently, the plugin failed momentarily (unrelated reasons) and we noticed that 50% of traffic was going to the canary service. After some investigation, we found out that, due to this issue, argoCD unset the weights we had setup to the services.
Environment:
- Contour version: 1.33.2
- Kubernetes version: (use
kubectl version): 1.33.8
- Kubernetes installer & version: N/A
- Cloud provider or hardware configuration: N/A
- OS (e.g. from
/etc/os-release): N/A
What steps did you take and what happened:
What did you expect to happen:
The weights are not managed by the manifest, so they should remain the same after step 3.
Anything else you would like to add:
We've been using this argo rollouts plugin to manage weights on our httpproxies. It changes the weights during canary rollouts gradually. Recently, the plugin failed momentarily (unrelated reasons) and we noticed that 50% of traffic was going to the canary service. After some investigation, we found out that, due to this issue, argoCD unset the weights we had setup to the services.
Environment:
kubectl version): 1.33.8/etc/os-release): N/A