Customer-facing docs for Pebble Sonar on Google Cloud Marketplace. Deploy on GKE, opt workloads into power profiling, and review results in the bundled UI.
oci://us-docker.pkg.dev/pebble-infra/sonar/sonar-v11.0)pebble-sonar-v1 (Helm CLI default)Install Sonar on your GKE cluster, then use the support page if you need help from Pebble.
Prerequisites, Helm install for NVIDIA and AMD clusters, Marketplace app-name/namespace commands, Hugging Face token setup, and post-install checks.
Read install guide → HelpSupport policy, contact channels, what to include with a request, and security reporting.
View support →Sonar discovers GPU inference workloads (vLLM, SGLang, generic GPU pods) on Kubernetes and, when a workload opts in, runs a power-profiler Job that measures throughput versus GPU power.
| Component | Purpose |
|---|---|
| Watcher | Discovers GPU workloads and schedules profiler Jobs when opted in |
| Prometheus | Scrapes DCGM (NVIDIA) or AMD device metrics exporter |
| Adminer UI | Browse SQLite discovery and profiler results |
| Profiler Jobs | Measure throughput at different GPU power levels (opt-in only) |