When exposing Nexus via NGINX Ingress on Kubernetes, users might see 413 request entity too large error

[ricardozanini]

Is your feature request related to a problem? Please describe.
When users deploy Nexus and try to mvn:deploy, a 413 error might happen since the default NGINX Ingress is not correctly configured. Would be nice to deploy a ConfigMap with it to address this problem and let users to adjust the request size as their needs. Example:

kind: ConfigMap
apiVersion: v1
metadata:
  name: nginx-configuration
  namespace: ingress-nginx
  labels:
    app.kubernetes.io/name: ingress-nginx
    app.kubernetes.io/part-of: ingress-nginx
data:
  proxy-body-size: 10m

Describe the solution you'd like
Add a ConfigMap to address this problem, the configuration should not be modified/reconciled by the Operator

Describe alternatives you've considered
A workaround is to use NodePort instead or create a custom Ingress instead.

Additional context
Might be related: kubernetes/ingress-nginx#4825

[ricardozanini]

@Kaitou786 FYI

[Kaitou786]

@ricardozanini I think we can have a small section in README about it. As this would highly depend which kind of ingress-controller was used. The above solution works with ingress nginx which is easily integrated with Minikube but not with other controllers. The data needs to be added in the configMap for the ingress controller which is end-user responsibility and choice and not the ingress object (I think beyond our scope). For the ingress controller deployed on Minikube they attach a configMap with it in the kube-system namespace usually named as nginx-load-balancer-conf``. Though this can be confirmed via running kubectl get deploy/ingress-nginx-controller -o yaml -n kube-system | grep configand checking the name in--configmap` flag.

[ricardozanini]

Have you managed to fix this error in your environment? Agreed to add a section in the README. The NGINX Ingress is the most common used on Kubernetes clusters, but not the only one, we can't cover all the scenarios. On the other hand, we can add a flag to the Networking interface for users to opt-in to use/indicate NGINX controller. I'll figure this out. We might be able to discover which Ingress type is being used and add this CM for users.

[Kaitou786]

Yes, I just tried adding this info in the configMap and it worked in my enviornment =D. Need to check if there is option to have no limit when uploading the file. I tried with 500m it didn't work so I gave 1500m and it started working
EDIT: Directly define 0m(for no limit) in the config map works just fine. I though it should be only 0. but that gave an error * patch: Invalid value: "map[data:map[proxy-body-size:0]]": cannot convert int64 to string but setting it to 0m worked.

[ricardozanini]

You might have to use m because of the Resource unit type. Whatever, glad that worked!

Perfect, then I'll add the CM if the Ingress is bound to a NGINX server.

[Kaitou786]

You might have to use m because of the Resource unit type. Whatever, glad that worked!

Perfect, then I'll add the CM if the Ingress is bound to a NGINX server.

Great!

[ricardozanini]

@Kaitou786 I'm considering to close this issue since it's a NGINX Ingress specific configuration, also the deployed CM on kube-system can address this configuration, right? The config you added was in this CM or you created a new one?

[Kaitou786]

Yes,the configuration was added to the CM that was created by the add-on in minikube. i just edited that to add the data field

[ricardozanini]

Ok, I'll close this one then.