What is the best practice for a multi-platform/Linux-based open source monitoring solution?

Question

Monitoring

Already tried:
ELK for logging (with filebeat),
Node_exporter (+ text-collector) + Prometeus + Grafana for metrics,
Telegraf + Influx + Grafana for metrics
But I still have no idea what tools professional DevOps use on their Kubernetis clusters for these purposes. New to DevOps (now learning Ansible, Jenkins, Docker). Therefore, please give advice if you have experience or know how it works in large companies.
In simple terms: I am currently looking for a solution to measure CPU temperature, like this. Heard that Zabbix/Nagios can do this out of the box, but is it necessary to have experience with each stack to make a feature rich monitoring system?
In your opinion, what is the best practice for an open source solution for multi-platform/Linux solutions?
I apologize for the offtopic, it's just that at the moment there is a mess in my head and there is a sale as an experience, but it is not structured at all. :) Maybe there are enlightening books on DevOps that you could recommend?
Sincerely

Answer 1

It seems to me that everyone has an individual stack. Research is carried out, the optimal solution is selected. We used:

• Zabbix (+Grafana) for metrics, monitoring and alerting
  
• ELK (+Grafana) for logging and alerting
  
• Graylog is currently under investigation for functional advantages/disadvantages over ELK
  

Answer 2

But I still have no idea what tools professional DevOps use on their Kubernetis clusters for these purposes.

The architecture of metrics and their processing in k8s . As a rule, Prometheus is used for unification. under it, a lot has already been sharpened in the cuber itself.
The servers themselves, on which the Kubernetes nodes are running, are monitored by the chosen solution for the entire infrastructure, be it Zabbix or some kind of Ionix.