Make rateAccuracy a setting

[akelge]

What would you like to be added: Currently when I configure metrics reading on a cluster I have 5 options plus autodetect. Each option has its own hardcoded value for rateAccuracy, I would like if this can become a setting parameter.

Why is this needed: I use kube-prometheus-stack helm chart to deploy the full stack on K8S clusters, but I would like to change the scrapeInterval to 1 minute instead of the default 30 seconds. If I do so this breaks CPU reporting in Lens, cause the detected prometheus (operator) has a default rateAccuracy of 1 minute that makes some CPU metrics query report empty results.

Environment you are Lens application on:

• Kubernetes distribution: not relevant
• Desktop OS: not relevant

[zalejus]

Hello everyone, January 26, 2024 and I am in exactly the same situation. Same problem - unable to change scraping interval in Prometheus without losing CPU usage in Lens panel. If you ask me it should be one of the basic functionalities.

[dark-brains]

Yes same issue not working CPU Usage metrics for example for Node