feat(support): add a redacted HAMi diagnostic bundle collector

.github/ISSUE_TEMPLATE/bug-report.md

@@ -18,9 +18,19 @@ labels: kind/bug
 
 Please include the following information:
 
+> **Preferred:** create a redacted archive first and inspect it before sharing:
+>
+> ```bash
+> hack/hami-support-bundle.sh --kubeconfig /path/to/config --context my-context \
+>   --namespace hami-system --pod affected-pod --output-dir ./hami-support
+> ```
+>
+> The collector never queries Kubernetes Secrets. Use `--help` for node and
+> cluster-wide options, and attach only artifacts appropriate for the issue.
+
 - The Pod manifest, or at least its accelerator resource requests and limits
 - Pod events and relevant accelerator-related node annotations
-- Output from the vendor's diagnostic tool, when available (for example, `nvidia-smi -a` for NVIDIA devices)
+- Output from the vendor's diagnostic tool, when available (for example, `nvidia-smi -q` for NVIDIA devices). The support-bundle script runs these optional tools only on the machine where the script is invoked; it does not run them on remote Kubernetes nodes.
 - Your container runtime configuration (for example, `/etc/docker/daemon.json` or the relevant containerd configuration)
 - Logs from the device plugin for the affected accelerator
 - The hami-scheduler logs

Makefile

@@ -72,6 +72,10 @@ test:
 bench:
 	bash hack/bench.sh
 
+.PHONY: test-support-bundle
+test-support-bundle:
+	bash hack/test-hami-support-bundle.sh
+
 lint:
 	bash hack/verify-staticcheck.sh