flux-metrics-api 0.0.1

Custom metrics exporter for Flux in Kubernetes

Flux Metrics API

This is an experiment to create a metrics API for Kubernetes that can be run directly from the Flux leader broker pod. We made this after creating prometheus-flux and wanting a more minimalist design. I'm not even sure it will work, but it's worth a try!

Usage

Install

You can install from pypi or from source:

$ python -m venv env
$ source env/bin/activate
$ pip install flux-metrics-api

# or

$ git clone https://github.com/converged-computing/flux-metrics-api
$ cd flux-metrics-api
$ pip install .
# you can also do "pip install -e ."

This will install the executable to your path, which might be your local user bin:

$ which flux-metric-api
/home/vscode/.local/bin/flux-metrics-api

Note that the provided .devcontainer includes an environment for VSCode where you have Flux and can install this and use ready to go!

Start

You'll want to be running in a Flux instance, as we need to connect to the broker handle.

$ flux start --test-size=4

And then start the server. This will use a default port and host (0.0.0.0:8443) that you can customize if desired.

$ flux-metrics-api start

# customize the port or host
$ flux-metrics-api start --port 9000 --host 127.0.0.1

If you want ssl (port 443) you can provide the path to a certificate and keyfile:

$ flux-metrics-api start --ssl-certfile /etc/certs/tls.crt --ssl-keyfile /etc/certs/tls.key

An example of a full command we might run from within a pod:

$ flux-metrics-api start --port 8443 --ssl-certfile /etc/certs/tls.crt --ssl-keyfile /etc/certs/tls.key --namespace flux-operator --service-name custom-metrics-apiserver

See --help to see other options available.

Endpoints

Metric

GET /apis/custom.metrics.k8s.io/v1beta2/namespaces//metrics/<metric_name>

Here is an example to get the "node_up_count" metric:

 curl -s http://localhost:8443/apis/custom.metrics.k8s.io/v1beta2/namespaces/flux-operator/metrics/node_up_count | jq

{
  "items": [
    {
      "metric": {
        "name": "node_up_count"
      },
      "value": 2,
      "timestamp": "2023-05-31T04:44:57+00:00",
      "windowSeconds": 0,
      "describedObject": {
        "kind": "Service",
        "namespace": "flux-operator",
        "name": "custom-metrics-apiserver",
        "apiVersion": "v1beta2"
      }
    }
  ],
  "apiVersion": "custom.metrics.k8s.io/v1beta2",
  "kind": "MetricValueList",
  "metadata": {
    "selfLink": "/apis/custom.metrics.k8s.io/v1beta2"
  }
}

The following metrics are supported:

• node_up_count: number of nodes up in the MiniCluster
• node_free_count: number of nodes free in the MiniCluster
• node_cores_free_count: number of node cores free in the MiniCluster
• node_cores_up_count: number of node cores up in the MiniCluster

Docker

We have a docker container, which you can customize for your use case, but it's more intended to be a demo. You can either build it yourself, or use our build.

$ docker build -t flux_metrics_api .
$ docker run -it -p 8443:8443 flux_metrics_api

or

$ docker run -it -p 8443:8443 ghcr.io/converged-computing/flux-metrics-api

Development

Note that this is implemented in Python, but (I found this after) we could also use Go. Specifically, I found this repository useful to see the spec format.

You can then open up the browser at http://localhost:8443/metrics/ to see the metrics!

😁️ Contributors 😁️

We use the all-contributors tool to generate a contributors graphic below.

Vanessasaurus 💻

License

HPCIC DevTools is distributed under the terms of the MIT license. All new contributions must be made under this license.

See LICENSE, COPYRIGHT, and NOTICE for details.

SPDX-License-Identifier: (MIT)

LLNL-CODE- 842614