diff --git a/applications/prompt-proto-service-lsstcomcamsim/values-usdfprod-prompt-processing.yaml b/applications/prompt-proto-service-lsstcomcamsim/values-usdfprod-prompt-processing.yaml
index 773d71ca4f..35131213a8 100644
--- a/applications/prompt-proto-service-lsstcomcamsim/values-usdfprod-prompt-processing.yaml
+++ b/applications/prompt-proto-service-lsstcomcamsim/values-usdfprod-prompt-processing.yaml
@@ -4,8 +4,8 @@ prompt-proto-service:
     # Expect to need roughly n_detector × request_latency / survey_cadence pods
     # For a 30 s ComCam survey with 500 s latency, this is 150
     # HACK: disable autoscaling as an OR4 workaround for DM-41829
-    autoscaling.knative.dev/min-scale: "100"
-    autoscaling.knative.dev/max-scale: "100"
+    autoscaling.knative.dev/min-scale: "150"
+    autoscaling.knative.dev/max-scale: "150"
     # Update this field if using latest or static image tag in dev
     revision: "1"