fix READMEs

Signed-off-by: Arvind Thirumurugan <arvindth@microsoft.com>

Author: Arvind Thirumurugan

approval-controller-metric-collector/approval-request-controller/README.md

@@ -55,24 +55,67 @@ helm install approval-request-controller ./charts/approval-request-controller \
 
 ## Configuration
 
-Key settings:
-- `controller.logLevel`: Log verbosity (default: 2)
-- `controller.resources`: Resource requests and limits
-- `rbac.create`: Create RBAC resources (default: true)
-- `crds.install`: Install MetricCollector, MetricCollectorReport, and WorkloadTracker CRDs (default: true)
-- `rbac.create`: Create RBAC resources (default: true)
-- `crds.install`: Install MetricCollector and MetricCollectorReport CRDs (default: true)
+The controller watches for:
+- `ApprovalRequest` (namespaced)
+- `ClusterApprovalRequest` (cluster-scoped)
 
-## Development
+Both resources from kubefleet are monitored, and the controller creates `MetricCollector` resources on appropriate member clusters based on the staged update configuration.
 
-### Build
+### Health Check Interval
 
-```bash
-make docker-build
+The controller checks workload health every **15 seconds**. This interval is configurable via the `reconcileInterval` parameter in the Helm chart.
+
+## API Reference
+
+### WorkloadTracker
+
+`WorkloadTracker` is a cluster-scoped custom resource that defines which workloads the approval controller should monitor for health metrics before auto-approving staged rollouts.
+
+#### Example: Single Workload
+
+```yaml
+apiVersion: placement.kubernetes-fleet.io/v1beta1
+kind: WorkloadTracker
+metadata:
+  name: sample-workload-tracker
+workloads:
+  - name: sample-metric-app
+    namespace: test-ns
 ```
 
-### Test Locally
+#### Example: Multiple Workloads
+
+```yaml
+apiVersion: placement.kubernetes-fleet.io/v1beta1
+kind: WorkloadTracker
+metadata:
+  name: multi-workload-tracker
+workloads:
+  - name: frontend
+    namespace: production
+  - name: backend-api
+    namespace: production
+  - name: worker-service
+    namespace: production
+```
 
-```bash
-go run ./cmd/approvalrequestcontroller/main.go
+#### Usage Notes
+
+- **Cluster-scoped:** WorkloadTracker is a cluster-scoped resource, not namespaced
+- **Optional:** If no WorkloadTracker exists, the controller will skip health checks and won't auto-approve
+- **Single instance:** The controller expects one WorkloadTracker per cluster and uses the first one found
+- **Health criteria:** All workloads listed must report healthy (metric value = 1.0) before approval
+- **Prometheus metrics:** Each workload should expose `workload_health` metrics that the MetricCollector can query
+
+For a complete example, see: [`./examples/workloadtracker/workloadtracker.yaml`](./examples/workloadtracker/workloadtracker.yaml)
+
+## Additional Resources
+
+- **Main Tutorial:** See [`../README.md`](../README.md) for a complete end-to-end tutorial on setting up automated staged rollouts with approval automation
+- **Metric Collector:** See [`../metric-collector/README.md`](../metric-collector/README.md) for details on the metric collection component that runs on member clusters
+- **KubeFleet Documentation:** [Azure/fleet](https://github.com/Azure/fleet) - Multi-cluster orchestration platform
+- **Example Configurations:**
+  - [`./examples/workloadtracker/`](./examples/workloadtracker/) - WorkloadTracker resource examples
+  - [`./examples/stagedupdaterun/`](./examples/stagedupdaterun/) - Staged update configuration examples
+  - [`./examples/prometheus/`](./examples/prometheus/) - Prometheus deployment and configuration for metric collection
 ```