Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Prediction analytics dashboard not capturing prediction API calls #168

Closed
aritrachatterjee15 opened this issue Jun 15, 2018 · 5 comments
Closed

Comments

@aritrachatterjee15
Copy link

aritrachatterjee15 commented Jun 15, 2018

I have been successful in deploying transformers and models in Seldon by following the Minikube with RBAC notebook. I am also able to successfully call the prediction API to get back scores. However, the API calls are not being logged by the Grafana dashboard.

I checked the prometheus logs and it is throwing out the following errors repeatedly:

time="2018-06-15T23:33:43Z" level=error msg="github.com/prometheus/prometheus/discovery/kubernetes/kubernetes.go:179: Failed to list *v1.Endpoints: endpoints is forbidden: User "system:serviceaccount:seldon:seldon" cannot list endpoints at the cluster scope" component="kube_client_runtime" source="kubernetes.go:75" 
time="2018-06-15T23:33:43Z" level=error msg="github.com/prometheus/prometheus/discovery/kubernetes/kubernetes.go:180: Failed to list *v1.Service: services is forbidden: User "system:serviceaccount:seldon:seldon" cannot list services at the cluster scope" component="kube_client_runtime" source="kubernetes.go:75" 
time="2018-06-15T23:33:43Z" level=error msg="github.com/prometheus/prometheus/discovery/kubernetes/kubernetes.go:181: Failed to list *v1.Pod: pods is forbidden: User "system:serviceaccount:seldon:seldon" cannot list pods at the cluster scope" component="kube_client_runtime" source="kubernetes.go:75" 

I wonder if prometheus deployment should be done using the prometheus service account available under the default namespace instead of using the seldon service account.

@ukclivecox
Copy link
Contributor

Can you tell us which commands/notebook you have followed to setup seldon-core? Also, how you created your kubernetes cluster?

This certainly seems like a missing/incorrect rolebinding in the RBAC setup.

@aritrachatterjee15
Copy link
Author

I followed the kubectl_demo_minikube_rbac notebook.

@ukclivecox
Copy link
Contributor

ukclivecox commented Jun 16, 2018

if you run:

kubectl create clusterrolebinding kube-system-cluster-admin-seldon --clusterrole=cluster-admin --serviceaccount=seldon:seldon

Does it fix it for you? If so we can update the notebook.

@aritrachatterjee15
Copy link
Author

Yes, that fixes the issue. Thanks!

@ukclivecox
Copy link
Contributor

There was an issue in the helm chart for the analytics setting up the serviceaccount. This should be fixed by #169.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants