init cortex placement/scheduling quality exporter.

cmd/scheduling-quality-exporter/main.go

@@ -0,0 +1,111 @@
+// Copyright SAP SE
+// SPDX-License-Identifier: Apache-2.0
+
+package main
+
+import (
+	"context"
+	"flag"
+	"fmt"
+	"log/slog"
+	"net/http"
+	"os"
+	"os/signal"
+	"syscall"
+	"time"
+
+	"github.com/prometheus/client_golang/prometheus"
+	"github.com/prometheus/client_golang/prometheus/promhttp"
+
+	"github.com/cobaltcore-dev/cortex/internal/exporters/schedulingquality/collector"
+	"github.com/cobaltcore-dev/cortex/internal/exporters/schedulingquality/nova"
+)
+
+func main() {
+	var (
+		listenAddr     = flag.String("listen-address", ":9199", "address to listen on for metrics")
+		metricsPath    = flag.String("metrics-path", "/metrics", "path under which to expose metrics")
+		novaEndpoint   = flag.String("nova-endpoint", "", "OpenStack Nova API endpoint (optional; reads OS_AUTH_URL env if unset)")
+		scrapeInterval = flag.Duration("scrape-interval", 30*time.Second, "minimum interval between Nova API calls")
+		kubeconfig     = flag.String("kubeconfig", "", "path to kubeconfig (uses in-cluster config if unset)")
+		isNovaDisabled = flag.Bool("disable-nova", false, "disable Nova flavor lookups; use fixed minimum placeable unit")
+		minCPU         = flag.Int64("min-cpu", 1, "minimum placeable CPU cores (used when Nova is disabled)")
+		minMemoryMB    = flag.Int64("min-memory-mb", 512, "minimum placeable memory in MB (used when Nova is disabled)")
+		logLevel       = flag.String("log-level", "info", "log level (debug, info, warn, error)")
+	)
+	flag.Parse()
+
+	level := slog.LevelInfo
+	switch *logLevel {
+	case "debug":
+		level = slog.LevelDebug
+	case "warn":
+		level = slog.LevelWarn
+	case "error":
+		level = slog.LevelError
+	}
+	logger := slog.New(slog.NewJSONHandler(os.Stderr, &slog.HandlerOptions{Level: level}))
+	slog.SetDefault(logger)
+
+	ctx, cancel := signal.NotifyContext(context.Background(), syscall.SIGINT, syscall.SIGTERM)
+	defer cancel()
+
+	var novaClient *nova.Client
+	if !*isNovaDisabled {
+		endpoint := *novaEndpoint
+		if endpoint == "" {
+			endpoint = os.Getenv("OS_AUTH_URL")
+		}
+		if endpoint != "" {
+			var err error
+			novaClient, err = nova.NewClientFromEnv()
+			handleError(err, "failed to create nova client")
+		}
+	}
+
+	k8sClient, err := collector.NewK8sClient(*kubeconfig)
+	handleError(err, "failed to create kubernetes client")
+
+	c := collector.New(collector.Options{
+		ScrapeInterval: *scrapeInterval,
+		NovaClient:     novaClient,
+		K8sClient:      k8sClient,
+		MinCPU:         *minCPU,
+		MinMemoryBytes: *minMemoryMB * 1024 * 1024,
+	})
+	prometheus.MustRegister(c)
+
+	mux := http.NewServeMux()
+	mux.Handle(*metricsPath, promhttp.Handler())
+	mux.HandleFunc("/healthz", func(w http.ResponseWriter, _ *http.Request) {
+		w.WriteHeader(http.StatusOK)
+		fmt.Fprintln(w, "ok")
+	})
+
+	srv := &http.Server{
+		Addr:         *listenAddr,
+		Handler:      mux,
+		ReadTimeout:  10 * time.Second,
+		WriteTimeout: 30 * time.Second,
+	}
+
+	go func() {
+		slog.Info("starting cortex-scheduling-quality-exporter", "address", *listenAddr, "path", *metricsPath)
+		if err := srv.ListenAndServe(); err != nil && err != http.ErrServerClosed {
+			handleError(err, "server error")
+		}
+	}()
+
+	<-ctx.Done()
+	slog.Info("shutting down")
+	shutdownCtx, shutdownCancel := context.WithTimeout(context.Background(), 5*time.Second)
+	defer shutdownCancel()
+	_ = srv.Shutdown(shutdownCtx)
+}
+
+func handleError(err error, message string) {
+	if err != nil {
+		slog.Error(message, "error", err)
+		os.Exit(1)
+	}
+}

helm/library/cortex-scheduling-quality-exporter/Chart.yaml

@@ -0,0 +1,17 @@
+apiVersion: v2
+name: cortex-scheduling-quality-exporter
+description: Prometheus exporter for Cortex scheduling quality and resource efficiency metrics
+type: application
+version: 0.1.0
+appVersion: "sha-9f115a57"
+keywords:
+  - prometheus
+  - exporter
+  - openstack
+  - cortex
+  - scheduling
+sources:
+  - https://github.com/cobaltcore-dev/cortex
+maintainers:
+  - name: cobaltcore-dev
+dependencies: []

helm/library/cortex-scheduling-quality-exporter/templates/_helpers.tpl

@@ -0,0 +1,72 @@
+{{/*
+Expand the name of the chart.
+*/}}
+{{- define "cortex-scheduling-quality-exporter.name" -}}
+{{- default .Chart.Name .Values.nameOverride | trunc 63 | trimSuffix "-" }}
+{{- end }}
+
+{{/*
+Create a default fully qualified app name.
+*/}}
+{{- define "cortex-scheduling-quality-exporter.fullname" -}}
+{{- if .Values.fullnameOverride }}
+{{- .Values.fullnameOverride | trunc 63 | trimSuffix "-" }}
+{{- else }}
+{{- $name := default .Chart.Name .Values.nameOverride }}
+{{- if contains $name .Release.Name }}
+{{- .Release.Name | trunc 63 | trimSuffix "-" }}
+{{- else }}
+{{- printf "%s-%s" .Release.Name $name | trunc 63 | trimSuffix "-" }}
+{{- end }}
+{{- end }}
+{{- end }}
+
+{{/*
+Create chart label.
+*/}}
+{{- define "cortex-scheduling-quality-exporter.chart" -}}
+{{- printf "%s-%s" .Chart.Name .Chart.Version | replace "+" "_" | trunc 63 | trimSuffix "-" }}
+{{- end }}
+
+{{/*
+Common labels.
+*/}}
+{{- define "cortex-scheduling-quality-exporter.labels" -}}
+helm.sh/chart: {{ include "cortex-scheduling-quality-exporter.chart" . }}
+{{ include "cortex-scheduling-quality-exporter.selectorLabels" . }}
+{{- if .Chart.AppVersion }}
+app.kubernetes.io/version: {{ .Chart.AppVersion | quote }}
+{{- end }}
+app.kubernetes.io/managed-by: {{ .Release.Service }}
+{{- end }}
+
+{{/*
+Selector labels.
+*/}}
+{{- define "cortex-scheduling-quality-exporter.selectorLabels" -}}
+app.kubernetes.io/name: {{ include "cortex-scheduling-quality-exporter.name" . }}
+app.kubernetes.io/instance: {{ .Release.Name }}
+{{- end }}
+
+{{/*
+ServiceAccount name.
+*/}}
+{{- define "cortex-scheduling-quality-exporter.serviceAccountName" -}}
+{{- default (include "cortex-scheduling-quality-exporter.fullname" .) .Values.serviceAccount.name }}
+{{- end }}
+
+{{/*
+OpenStack credentials Secret name.
+*/}}
+{{- define "cortex-scheduling-quality-exporter.osSecretName" -}}
+{{- include "cortex-scheduling-quality-exporter.fullname" . }}-openstack
+{{- end }}
+
+{{/*
+Returns true if Nova integration is enabled (all required openstack fields are set).
+*/}}
+{{- define "cortex-scheduling-quality-exporter.novaEnabled" -}}
+{{- if and .Values.openstack.authURL .Values.openstack.username .Values.openstack.password .Values.openstack.projectName -}}
+true
+{{- end }}
+{{- end }}

helm/library/cortex-scheduling-quality-exporter/templates/clusterrole.yaml

@@ -0,0 +1,30 @@
+apiVersion: rbac.authorization.k8s.io/v1
+kind: ClusterRole
+metadata:
+  name: {{ include "cortex-scheduling-quality-exporter.fullname" . }}
+  labels:
+    {{- include "cortex-scheduling-quality-exporter.labels" . | nindent 4 }}
+rules:
+  - apiGroups:
+      - openstack.sapcloud.io
+    resources:
+      - hypervisors
+    verbs:
+      - get
+      - list
+      - watch
+---
+apiVersion: rbac.authorization.k8s.io/v1
+kind: ClusterRoleBinding
+metadata:
+  name: {{ include "cortex-scheduling-quality-exporter.fullname" . }}
+  labels:
+    {{- include "cortex-scheduling-quality-exporter.labels" . | nindent 4 }}
+roleRef:
+  apiGroup: rbac.authorization.k8s.io
+  kind: ClusterRole
+  name: {{ include "cortex-scheduling-quality-exporter.fullname" . }}
+subjects:
+  - kind: ServiceAccount
+    name: {{ include "cortex-scheduling-quality-exporter.serviceAccountName" . }}
+    namespace: {{ .Release.Namespace }}

helm/library/cortex-scheduling-quality-exporter/templates/deployment.yaml

@@ -0,0 +1,79 @@
+apiVersion: apps/v1
+kind: Deployment
+metadata:
+  name: {{ include "cortex-scheduling-quality-exporter.fullname" . }}
+  labels:
+    {{- include "cortex-scheduling-quality-exporter.labels" . | nindent 4 }}
+spec:
+  replicas: {{ .Values.replicaCount }}
+  selector:
+    matchLabels:
+      {{- include "cortex-scheduling-quality-exporter.selectorLabels" . | nindent 6 }}
+  template:
+    metadata:
+      annotations:
+        prometheus.io/scrape: "true"
+        prometheus.io/port: {{ .Values.port | quote }}
+        prometheus.io/path: {{ .Values.config.metricsPath | quote }}
+        {{- with .Values.podAnnotations }}
+        {{- toYaml . | nindent 8 }}
+        {{- end }}
+      labels:
+        {{- include "cortex-scheduling-quality-exporter.labels" . | nindent 8 }}
+    spec:
+      {{- with .Values.imagePullSecrets }}
+      imagePullSecrets:
+        {{- toYaml . | nindent 8 }}
+      {{- end }}
+      serviceAccountName: {{ include "cortex-scheduling-quality-exporter.serviceAccountName" . }}
+      securityContext:
+        {{- toYaml .Values.podSecurityContext | nindent 8 }}
+      containers:
+        - name: exporter
+          securityContext:
+            {{- toYaml .Values.securityContext | nindent 12 }}
+          image: "{{ .Values.image.repository }}:{{ .Values.image.tag | default .Chart.AppVersion }}"
+          imagePullPolicy: {{ .Values.image.pullPolicy }}
+          args:
+            - --listen-address=:{{ .Values.port }}
+            - --metrics-path={{ .Values.config.metricsPath }}
+            - --scrape-interval={{ .Values.config.scrapeInterval }}
+            - --log-level={{ .Values.config.logLevel }}
+            {{- if not (include "cortex-scheduling-quality-exporter.novaEnabled" .) }}
+            - --disable-nova
+            {{- end }}
+          {{- if include "cortex-scheduling-quality-exporter.novaEnabled" . }}
+          envFrom:
+            - secretRef:
+                name: {{ include "cortex-scheduling-quality-exporter.osSecretName" . }}
+          {{- end }}
+          ports:
+            - name: metrics
+              containerPort: {{ .Values.port }}
+              protocol: TCP
+          livenessProbe:
+            httpGet:
+              path: /healthz
+              port: metrics
+            initialDelaySeconds: 5
+            periodSeconds: 30
+          readinessProbe:
+            httpGet:
+              path: /healthz
+              port: metrics
+            initialDelaySeconds: 5
+            periodSeconds: 15
+          resources:
+            {{- toYaml .Values.resources | nindent 12 }}
+      {{- with .Values.nodeSelector }}
+      nodeSelector:
+        {{- toYaml . | nindent 8 }}
+      {{- end }}
+      {{- with .Values.affinity }}
+      affinity:
+        {{- toYaml . | nindent 8 }}
+      {{- end }}
+      {{- with .Values.tolerations }}
+      tolerations:
+        {{- toYaml . | nindent 8 }}
+      {{- end }}

helm/library/cortex-scheduling-quality-exporter/templates/secret.yaml

@@ -0,0 +1,16 @@
+{{- if include "cortex-scheduling-quality-exporter.novaEnabled" . }}
+apiVersion: v1
+kind: Secret
+metadata:
+  name: {{ include "cortex-scheduling-quality-exporter.osSecretName" . }}
+  labels:
+    {{- include "cortex-scheduling-quality-exporter.labels" . | nindent 4 }}
+type: Opaque
+stringData:
+  OS_AUTH_URL: {{ .Values.openstack.authURL | quote }}
+  OS_USERNAME: {{ .Values.openstack.username | quote }}
+  OS_PASSWORD: {{ .Values.openstack.password | quote }}
+  OS_PROJECT_NAME: {{ .Values.openstack.projectName | quote }}
+  OS_USER_DOMAIN_NAME: {{ .Values.openstack.userDomainName | quote }}
+  OS_PROJECT_DOMAIN_NAME: {{ .Values.openstack.projectDomainName | quote }}
+{{- end }}

helm/library/cortex-scheduling-quality-exporter/templates/service.yaml

@@ -0,0 +1,15 @@
+apiVersion: v1
+kind: Service
+metadata:
+  name: {{ include "cortex-scheduling-quality-exporter.fullname" . }}
+  labels:
+    {{- include "cortex-scheduling-quality-exporter.labels" . | nindent 4 }}
+spec:
+  type: {{ .Values.service.type }}
+  ports:
+    - port: {{ .Values.port }}
+      targetPort: metrics
+      protocol: TCP
+      name: metrics
+  selector:
+    {{- include "cortex-scheduling-quality-exporter.selectorLabels" . | nindent 4 }}

helm/library/cortex-scheduling-quality-exporter/templates/serviceaccount.yaml

@@ -0,0 +1,10 @@
+apiVersion: v1
+kind: ServiceAccount
+metadata:
+  name: {{ include "cortex-scheduling-quality-exporter.serviceAccountName" . }}
+  labels:
+    {{- include "cortex-scheduling-quality-exporter.labels" . | nindent 4 }}
+  {{- with .Values.serviceAccount.annotations }}
+  annotations:
+    {{- toYaml . | nindent 4 }}
+  {{- end }}

helm/library/cortex-scheduling-quality-exporter/templates/servicemonitor.yaml

@@ -0,0 +1,21 @@
+{{- if .Values.serviceMonitor.enabled }}
+apiVersion: monitoring.coreos.com/v1
+kind: ServiceMonitor
+metadata:
+  name: {{ include "cortex-scheduling-quality-exporter.fullname" . }}
+  namespace: {{ .Values.serviceMonitor.namespace | default .Release.Namespace }}
+  labels:
+    {{- include "cortex-scheduling-quality-exporter.labels" . | nindent 4 }}
+    {{- with .Values.serviceMonitor.labels }}
+    {{- toYaml . | nindent 4 }}
+    {{- end }}
+spec:
+  selector:
+    matchLabels:
+      {{- include "cortex-scheduling-quality-exporter.selectorLabels" . | nindent 6 }}
+  endpoints:
+    - port: metrics
+      path: {{ .Values.config.metricsPath }}
+      interval: {{ .Values.serviceMonitor.interval }}
+      scrapeTimeout: {{ .Values.serviceMonitor.scrapeTimeout }}
+{{- end }}