Tools to analyze CPU and memory usage k8s metrics

[jitub]

Is there a tool in GKE/K8s that can automatically analyze CPU and memory usage metrics from Prometheus or any other metrics based data and provide recommendations to developers via an API, reducing the need for manual analysis and copy-pasting of resource requests/limits?

My developers are copy pasting cpu memory and every time I have to analyze the cpu memory metrics and give them recommendation based on past usage. I would like to see how I can automate this without writing some code.

[DarwinVinoth]

GCP Recommendation API
Google Cloud Recommender provides actionable recommendations for optimizing your Google Cloud resources, such as Compute Engine instances, but it currently does not offer dedicated recommendations for Kubernetes pod resource requests.

Use Case:
It can help with optimizing node pool configurations and other GCP resources, and you can integrate it via its REST API.

Ref: https://cloud.google.com/recommender/docs/use-api

VPA in Kuberenetes:

VPA continuously monitors pod CPU and memory usage, providing recommendations for resource requests and limits. In recommendation mode, it writes suggested values into the resource's status without automatically changing them.
Use Case:
You can query these recommendations via the Kubernetes API to guide developers in setting appropriate resource requests.

[mokit]

Hi, @jitub.

Before answering your question, could you please tell me which auto-scaling method you're using for the pods - HPA, VPA or both? Also, what monitoring tools are you currently using?

Regards,
Mokit

[jitub]

1. HPA.

2. Prometheus. 

[jayeshmahajan · Accepted solution]

KRR is good to get those https://github.com/robusta-dev/krr
Goldilocks is another interesting one that provides APIs to automate this.