kubernetes-sigs / dashboard-metrics-scraper

Container to scrape, store, and retrieve a window of time from the Metrics Server.
Apache License 2.0
87 stars 39 forks source link

Does not scrape any node or pods #17

Closed Vrtak-CZ closed 5 years ago

Vrtak-CZ commented 5 years ago

Hi,

i'm just deployed Kubernetes Dashboard 2.0.0-beta1 with this scraper 1.0.0 and in logs i see Database updated: 0 nodes, 0 pods but there are 3 nodes and lots of pods. And yes metrics-server is deployed in cluster. It's AWS EKS 1.13.7 cluster. Maybe i'm just doing something wrong.

For deployment i used this template: https://github.com/kubernetes/dashboard/blob/v2.0.0-beta1/aio/deploy/recommended.yaml

jeefy commented 5 years ago

Two things that come to mind.

Is there anything else in the scraper or metrics-server logs? Also, could you possibly do something like kubectl proxy and try to hit the metrics.k8s.io endpoint yourself? If you need any guidance with that, just ping me. :)

Thanks!

Vrtak-CZ commented 5 years ago

Thank you. So i made some research and debugging and problem is with metrics server resp. wit whole cluster because DNS in cluster is not able resolve *.*.compute.internal domains. So my problem is not related to metrics-scraper. One more time thank you.