Scale Your Own Way, Using HPA with Built-In Support for PromQL Metrics Queries in GKE
Google, Wednesday, September 23rd, 2026
GKE's Horizontal Pod Autoscaler now natively supports PromQL queries against Cloud Monitoring/Prometheus, no third-party adapters needed.
Google Kubernetes Engine now offers built-in support for processing Prometheus metrics with expressive PromQL queries to drive Horizontal Pod Autoscaler (HPA) triggers, building on earlier native custom-metrics support that cut metrics latency to five seconds.
HPA can now directly read autoscaling metrics from Cloud Monitoring via Google Managed Service for Prometheus without third-party adapters.
Google says the feature helps cases like scaling a worker pool on unacknowledged Pub/Sub messages or scaling an inference service on queries-per-second. Support for self-hosted Prometheus servers is planned as the feature moves toward general availability.