move scaling metric to a separate file for clarity

fastmachinelearning · Feb 7, 2025 · be05178 · be05178
1 parent 057c529
commit be05178
Show file tree

Hide file tree

Showing 3 changed files with 24 additions and 15 deletions.
diff --git a/helm/supersonic/templates/NOTES.txt b/helm/supersonic/templates/NOTES.txt
@@ -18,7 +18,7 @@ SuperSONIC chart successfully installed!
 
 Scaling metric:{{ if not ( eq .Values.prometheus.serverLoadMetric "" ) }}    {{ .Values.prometheus.serverLoadMetric }}{{ else }}{{ include "supersonic.defaultMetric" . | nindent 4 }}{{ end }}
 
-Scaling threshold: {{ .Values.prometheus.serverLoadThreshold }}{{"\n"}}
+Scaling threshold: {{ include "supersonic.serverLoadThreshold" . }}{{"\n"}}
 
 {{- if or (and .Values.grafana.enabled (eq (include "supersonic.grafanaExists" .) "true")) (eq (include "supersonic.prometheusExists" .) "true") .Values.prometheus.external }}
 ┌-----------------------------------------------------------------------------┐

diff --git a/helm/supersonic/templates/_helpers.tpl b/helm/supersonic/templates/_helpers.tpl
@@ -25,20 +25,6 @@ Get Envoy name
 {{- printf "%s-envoy" (include "supersonic.name" .) | trunc 63 | trimSuffix "-" -}}
 {{- end -}}
 
-{{- define "supersonic.defaultMetric" -}}
-{{- if not ( eq .Values.prometheus.serverLoadMetric "" ) }}
-  {{- printf "%s" .Values.prometheus.serverLoadMetric -}}
-{{- else }}
-sum by (release) (
-    rate(nv_inference_queue_duration_us{release=~"{{ include "supersonic.name" . }}"}[15s])
-)
-  /
-sum by (release) (
-    (rate(nv_inference_exec_count{release=~"{{ include "supersonic.name" . }}"}[15s]) * 1000) + 0.001
-)
-{{- end }}
-{{- end }}
-
 {{/*
 Get gRPC endpoint
 */}}

diff --git a/helm/supersonic/templates/_scaling-metric.tpl b/helm/supersonic/templates/_scaling-metric.tpl
@@ -0,0 +1,23 @@
+{{/*
+Get default scaling metric
+*/}}
+{{- define "supersonic.defaultMetric" -}}
+{{- if not ( eq .Values.prometheus.serverLoadMetric "" ) }}
+  {{- printf "%s" .Values.prometheus.serverLoadMetric -}}
+{{- else }}
+sum by (release) (
+    rate(nv_inference_queue_duration_us{release=~"{{ include "supersonic.name" . }}"}[15s])
+)
+  /
+sum by (release) (
+    (rate(nv_inference_exec_count{release=~"{{ include "supersonic.name" . }}"}[15s]) * 1000) + 0.001
+)
+{{- end }}
+{{- end }}
+
+{{/*
+Get server load threshold (defaults to 100 if not set)
+*/}}
+{{- define "supersonic.serverLoadThreshold" -}}
+{{- default 100 .Values.prometheus.serverLoadThreshold -}}
+{{- end -}}