I figure out that it is base64 encoded ssh private key on master node.
The monitoring stack still can not be deployed. For anyone encountering this issue, I share some investigations.
I looked the log at /var/lib/waagent/custom-script/download/0/ and source code at config.sh, found that the controller machine failed to get ./kube/config from master node
For quick solution, I copy ./kube/config manually from master node to controller machine's root account.
Execute manual_config.sh. At this step, you may encounter the line ending issue in config.sh file, you can change the line ending to LF instead of CRLF .
Finally, the monitoring is deployed successfully.
The default configuration of monitoring stack won't work because lack of permission if your cluster has enabled RBAC, you have to add serviceaccount to those pod according to Using RBAC Authorization
If kibana can not search any log, there might be a problem in the index configuration. You have to see the Parsing Logs with Logstash and modify the configmap of logstash
I am confused about this term. It is in /monitoring/readme
Is it a base64 encoded string where I can get from azure portal?
Or Is it a base64 encoded string that I have to generate on my master node machine? If yes, then how can I generate the identity file?