Logging and Monitoring
Summary
Monitor Cluster Component
There is a lot of open source plugin metrics like datadog, dynatrace or prometheuse... to get metrics like number of pods, cpu, ram...
Heapster is deprecated and now we can use metrics server instead. Metric server is an in-memory solution. kublet is reponsible to send informations and metrics about each node.
to add metrics server :
git clone https://github.com/kubernetes-incubator/metrics-serve kubectl create -f deploy/1.8+/ kubectl top node (to get nodes metrics) kubectl top pod (to get pods metrics)
sdds
Managing Application Logs
If you have only pods runned with docker, you can use this command :
docker logs -f ecf
In Kubernetes, the command is :
kubectl logs -f event-pod
If we have multi-container in a pod, we should specifie the name of the container
kubectl logs -f name-pod name-container