openfero

OpenFero is intended as an event-triggered job scheduler framework for code agnostic recovery jobs.

View project on GitHub

OpenFero

Open Fero is a little play on words from the Latin “opem fero”, which means “to help” and the term “OpenSource”. Hence the name “openfero”. The scope of OpenFero is a framework for self-healing in a cloud-native environment.

OpenSSF Scorecard OpenSSF Best Practices

Local development

export KUBE_VERSION=v1.26.0
export PROM_OPERATOR_VERSION=45.2.0

brew install kind
kind create cluster --image kindest/node:${KUBE_VERSION}
helm install mmop prometheus-community/kube-prometheus-stack --namespace default --set kubeTargetVersionOverride="${KUBE_VERSION}" --version=${PROM_OPERATOR_VERSION}

build image

make build

Test-Case

curl -X POST -H "Content-Type: application/json" -d @./test/alerts.json http://localhost:8080/alerts

Component-Diagram

Shows the Prometheus, Alertmanager components and that Alertmanager notifies the OpenFero component so that OpenFero starts the jobs via Kubernetes API.

Job definitions

The job definitions are stored in the namespace in ConfigMaps with the naming convention openfero-<alertname>-<status>.

Example-Names

  • openfero-KubeQuotaAlmostReached-firing
  • openfero-KubeQuotaAlmostReached-resolved

Job-Example

apiVersion: batch/v1
kind: Job
metadata:
  name: openfero-kubequotaalmostfull-firing
  labels:
    app: openfero
spec:
  parallelism: 1
  completions: 1
  template:
      labels:
        app: openfero
      spec:
        containers:
        - name: python-job
          image: python:latest
          args:
          - bash
          - -c
          - |-
            echo "Hallo Welt"
        imagePullPolicy: Always
        restartPolicy: Never
        serviceAccount: <desired-sa>
        serviceAccountName: <desired-sa>

Security note

The service account that is installed when deploying openfero is for openfero itself. For the jobs, separate service accounts must be rolled out, which have the appropriate permissions for the job.

For jobs that need to interact with the Kubernetes API, it is recommended to define a suitable role for the job and authorize it via ServiceAccount in the job definition.