Fetch logs from Elastic Search

[troeger]

The future version of our cluster will put logs into an Elastic Search database through Fluentd. Kubeportal would have a unique feature by fetching log streams for pods / namespaces from such an Elastic Search endpoint, and not the API server.

This issue is intended for a first discussion. I envision the following features:

• Fluentd marks log entries in Elastic Search with the K8S namespace and K8s pod name creating the log.
• The Kubeportal backend can be configured with an Elastic Search IP / user name / password, and with the names of the labels for namespace and pod.
• A new backend API endpoint allows to fetch a stream of logs for a pod. Every request to the Kubeportal backend endpoint leads to an ES query, so the Kubeportal API is just wrapping the log database.
• The new frontend can show the logs for a pod / deployment, or even complete namespace.

Later, we could add advanced filtering capabilities in the frontend. Another open question is if we want to implement this with WebSockets from the very start.