feat: add KEDA autoscaling support for OpenTelemetry Ingest with configurable metrics

HelmChart/Public/oneuptime/templates/_helpers.tpl

@@ -701,3 +701,62 @@ spec:
     requests:
       storage: {{ $.Storage }}
 {{- end }}
+
+
+{{/*
+KEDA ScaledObject template for metric-based autoscaling
+Usage: include "oneuptime.kedaScaledObject" (dict "ServiceName" "service-name" "Release" .Release "Values" .Values "MetricsConfig" {...})
+*/}}
+{{- define "oneuptime.kedaScaledObject" }}
+{{- if and .Values.keda.enabled .MetricsConfig.enabled (not .DisableAutoscaler) }}
+apiVersion: keda.sh/v1alpha1
+kind: ScaledObject
+metadata:
+  name: {{ printf "%s-%s-scaledobject" .Release.Name .ServiceName }}
+  namespace: {{ .Release.Namespace }}
+  labels:
+    app: {{ printf "%s-%s" .Release.Name .ServiceName }}
+    app.kubernetes.io/part-of: oneuptime
+    app.kubernetes.io/managed-by: Helm
+    appname: oneuptime
+spec:
+  scaleTargetRef:
+    name: {{ printf "%s-%s" .Release.Name .ServiceName }}
+  minReplicaCount: {{ .MetricsConfig.minReplicas }}
+  maxReplicaCount: {{ .MetricsConfig.maxReplicas }}
+  pollingInterval: {{ .MetricsConfig.pollingInterval }}
+  cooldownPeriod: {{ .MetricsConfig.cooldownPeriod }}
+  triggers:
+    {{- range .MetricsConfig.triggers }}
+    - type: prometheus
+      metadata:
+        serverAddress: http://{{ printf "%s-%s" $.Release.Name $.ServiceName }}:{{ .port }}/metrics
+        query: {{ .query }}
+        threshold: {{ .threshold | quote }}
+      authenticationRef:
+        name: {{ printf "%s-%s-trigger-auth" $.Release.Name $.ServiceName }}
+    {{- end }}
+---
+apiVersion: keda.sh/v1alpha1
+kind: TriggerAuthentication
+metadata:
+  name: {{ printf "%s-%s-trigger-auth" .Release.Name .ServiceName }}
+  namespace: {{ .Release.Namespace }}
+  labels:
+    app: {{ printf "%s-%s" .Release.Name .ServiceName }}
+    app.kubernetes.io/part-of: oneuptime
+    app.kubernetes.io/managed-by: Helm
+    appname: oneuptime
+spec:
+  secretTargetRef:
+    {{- if .Values.oneuptimeSecret }}
+    - parameter: X-Cluster-Key
+      name: {{ printf "%s-%s" .Release.Name "secrets" }}
+      key: oneuptime-secret
+    {{- else if .Values.externalSecrets.oneuptimeSecret.existingSecret.name }}
+    - parameter: X-Cluster-Key
+      name: {{ .Values.externalSecrets.oneuptimeSecret.existingSecret.name }}
+      key: {{ .Values.externalSecrets.oneuptimeSecret.existingSecret.passwordKey }}
+    {{- end }}
+{{- end }}
+{{- end }}

HelmChart/Public/oneuptime/templates/open-telemetry-ingest.yaml

@@ -112,8 +112,20 @@ spec:
 ---
 
 # OneUptime open-telemetry-ingest autoscaler
-{{- if not $.Values.openTelemetryIngest.disableAutoscaler }}
+{{- if and (not $.Values.openTelemetryIngest.disableAutoscaler) (not (and $.Values.keda.enabled $.Values.openTelemetryIngest.keda.enabled)) }}
 {{- $openTelemetryIngestAutoScalerArgs := dict "ServiceName" "open-telemetry-ingest" "Release" $.Release "Values" $.Values -}}
 {{- include "oneuptime.autoscaler" $openTelemetryIngestAutoScalerArgs }}
 {{- end }}
 ---
+
+{{/*
+KEDA ScaledObjects for various services
+*/}}
+
+{{/* OpenTelemetry Ingest KEDA ScaledObject */}}
+{{- if and .Values.keda.enabled .Values.openTelemetryIngest.keda.enabled (not .Values.openTelemetryIngest.disableAutoscaler) }}
+{{- $metricsConfig := dict "enabled" .Values.openTelemetryIngest.keda.enabled "minReplicas" .Values.openTelemetryIngest.keda.minReplicas "maxReplicas" .Values.openTelemetryIngest.keda.maxReplicas "pollingInterval" .Values.openTelemetryIngest.keda.pollingInterval "cooldownPeriod" .Values.openTelemetryIngest.keda.cooldownPeriod "triggers" (list (dict "query" "oneuptime_telemetry_queue_size" "threshold" .Values.openTelemetryIngest.keda.queueSizeThreshold "port" .Values.port.openTelemetryIngest)) }}
+{{- $openTelemetryIngestKedaArgs := dict "ServiceName" "open-telemetry-ingest" "Release" .Release "Values" .Values "MetricsConfig" $metricsConfig "DisableAutoscaler" .Values.openTelemetryIngest.disableAutoscaler }}
+{{- include "oneuptime.kedaScaledObject" $openTelemetryIngestKedaArgs }}
+{{- end }}
+---

HelmChart/Public/oneuptime/values.yaml

@@ -496,6 +496,17 @@ openTelemetryIngest:
   disableTelemetryCollection: false
   disableAutoscaler: false
   resources:
+  # KEDA autoscaling configuration based on queue metrics
+  keda:
+    enabled: false
+    minReplicas: 1
+    maxReplicas: 100
+    # Scale up when queue size exceeds this threshold
+    queueSizeThreshold: 100
+    # Polling interval for metrics (in seconds)
+    pollingInterval: 30
+    # Cooldown period after scaling (in seconds)
+    cooldownPeriod: 300
 
 fluentIngest:
   replicaCount: 1