You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I followed the instructions exactly to make a REST model. I wrapped the example(ModelWithMetrics.py) with s2i seldonio/seldon-core-s2i-python3:0.7 and deployed the model on seldon-core to k8s.
However, when looking at the prometheus targets page I get an error: Get "http://[IP address]:6000/prometheus": dial tcp 10.0.2.85:6000: connect: connection refused
@jlee3095 this functionality was only introduced from v1.1.0+, you would have to re-wrap your container with a python wrapper v1.1.0 for the prometheus metrics endpoint is available from the model wrapper containers. Versions prior to 1.1.0 exposed metrics only through the service orchestrator.
I am trying to write custom metrics so that I can expose them on prometheus and grafana following this guide (https://docs.seldon.io/projects/seldon-core/en/v0.3.0/examples/tmpl_model_with_metrics.html) to create a model with custom metrics.
I followed the instructions exactly to make a REST model. I wrapped the example(ModelWithMetrics.py) with s2i seldonio/seldon-core-s2i-python3:0.7 and deployed the model on seldon-core to k8s.
However, when looking at the prometheus targets page I get an error:
Get "http://[IP address]:6000/prometheus": dial tcp 10.0.2.85:6000: connect: connection refused
In addition, I tried deploying the exact same already-wrapped model with custom metrics found in the github repo (https://raw.githubusercontent.com/SeldonIO/seldon-core/master/examples/models/custom_metrics/model_rest.yaml) and the 6000 port did not have any connection issues.
Any help would be appreciated.
The text was updated successfully, but these errors were encountered: