fix(newrelic-logging): correct logLevel default to enable global.verboseLog inheritance

- Changed fluentBit.logLevel default from "info" to "" (empty)
- This allows global.verboseLog=true to correctly set LOG_LEVEL=debug
- Added clear precedence comments in values.yaml
- Restored functional E2E test files (test-global-values.sh, README.md)

Root cause: values.yaml had logLevel: "info" as default, causing the
template's if $logLevel condition to always be true, preventing
global.verboseLog from being evaluated.

Test Results: All 107 helm-unittest tests pass
Template validation:
- Default (no settings) → LOG_LEVEL="info"
- global.verboseLog=true → LOG_LEVEL="debug"
- Explicit fluentBit.logLevel → takes precedence

Author: dpacheconr

charts/newrelic-logging/tests/functional-e2e/README.md

@@ -0,0 +1,371 @@
+# Functional E2E Tests for Global Value Inheritance
+
+## Overview
+
+This directory contains functional end-to-end (E2E) tests that validate global value inheritance for the newrelic-logging chart in a real Kubernetes environment. Unlike helm-unittest tests (which only validate template rendering), these tests deploy the chart to a Kubernetes cluster and verify that configuration values actually reach the running containers.
+
+**Important**: These tests do NOT validate telemetry (whether data reaches New Relic). They only validate functional configuration propagation. For telemetry validation, see the main repository E2E test suite using `newrelic-integration-e2e-action`.
+
+## What These Tests Validate
+
+The test suite validates that the following global values propagate correctly to the newrelic-logging DaemonSet:
+
+1. **Proxy Configuration** (`global.proxy`)
+   - HTTP_PROXY and HTTPS_PROXY environment variables are set in Fluent Bit containers
+   - Local `proxy` value overrides `global.proxy` (precedence validation)
+
+2. **NodeSelector** (`global.nodeSelector`)
+   - Global node selection constraints apply to DaemonSet pods
+   - Pods schedule only on nodes matching the selector
+
+3. **Tolerations** (`global.tolerations`)
+   - Global tolerations allow pods to tolerate node taints
+   - Pods can schedule on tainted nodes
+
+4. **Custom Registry** (`global.images.registry`)
+   - Container images pull from custom registry
+   - Image paths include the custom registry domain
+
+5. **ServiceAccount Annotations** (`global.serviceAccount.annotations`)
+   - Annotations (e.g., for IRSA, Workload Identity) propagate to ServiceAccount
+   - Critical for cloud IAM role bindings
+
+6. **Verbose Logging** (`global.verboseLog`)
+   - `global.verboseLog: true` maps to `LOG_LEVEL=debug` in containers
+   - Enables debug logging across all components
+
+7. **Host Network** (`global.hostNetwork`)
+   - `global.hostNetwork: true` enables host network mode for DaemonSet pods
+   - Required for certain networking configurations
+
+## Test Approach and Limitations
+
+### What These Tests Validate
+
+These tests validate **configuration propagation** - they verify that Helm values reach the Kubernetes pod specifications correctly. The tests:
+
+- ✅ Deploy the chart using Helm with specific global values
+- ✅ Wait for pod objects to be created in Kubernetes
+- ✅ Inspect pod specifications to verify configuration is present
+- ✅ Validate environment variables, nodeSelectors, tolerations, etc.
+
+### What These Tests Do NOT Validate
+
+These tests do NOT validate runtime behavior:
+
+- ❌ Pods do NOT need to become "Ready" (tests don't wait for readiness)
+- ❌ Fluent Bit does NOT need to actually run
+- ❌ No telemetry validation (data reaching New Relic)
+- ❌ No functional testing of log collection
+
+### Rationale
+
+This approach allows fast, lightweight testing of configuration inheritance without requiring:
+- Working container images in the test environment
+- New Relic account credentials
+- Long test execution times (tests complete in ~2-3 minutes vs ~10+ minutes)
+
+The tests use lightweight test images (or image pull policy `Never`) to speed up execution and reduce dependencies. Since we're only validating pod specifications (which Kubernetes populates immediately), we don't need the containers to successfully start.
+
+**For runtime and telemetry validation**, see the main repository E2E test suite using `newrelic-integration-e2e-action`.
+
+## Prerequisites
+
+### Required Tools
+
+- **kubectl**: Kubernetes command-line tool
+- **helm**: Helm 3.x
+- **Kubernetes cluster**: Minikube, KIND, or any Kubernetes cluster
+
+### Cluster Requirements
+
+- Cluster must have at least 1 node
+- User must have permissions to:
+  - Deploy DaemonSets and ServiceAccounts
+  - Label nodes
+  - Taint nodes
+  - Create/delete Helm releases
+
+### Test Environment Setup
+
+The tests are designed to run on local Kubernetes clusters (Minikube, KIND, k3d). Example setup with k3d:
+
+```bash
+# Install k3d (lightweight Kubernetes)
+brew install k3d  # macOS
+# OR
+# curl -s https://raw.githubusercontent.com/k3d-io/k3d/main/install.sh | bash
+
+# Create a local cluster
+k3d cluster create demo
+
+# Verify cluster is running
+kubectl get nodes
+
+# Build lightweight test image
+docker build -t e2e/newrelic-fluentbit-output:test -f Dockerfile.test .
+
+# Load image into k3d cluster
+k3d image import e2e/newrelic-fluentbit-output:test -c demo
+```
+
+**Note**: The test image can be any lightweight image (busybox, alpine) since tests only validate pod specifications, not runtime behavior. The chart is configured with `image.pullPolicy=Never` to use locally loaded images.
+
+## Running the Tests
+
+### Run All Tests
+
+```bash
+cd charts/newrelic-logging/tests/functional-e2e
+./test-global-values.sh
+```
+
+### Test Output
+
+The test script provides colored output:
+- 🟢 **[INFO]**: General information
+- 🟡 **[WARN]**: Warnings (non-fatal)
+- 🔴 **[ERROR]**: Errors (fatal)
+- ✓ **PASS**: Test passed
+- ✗ **FAIL**: Test failed
+
+Example output:
+
+```
+[INFO] Starting functional E2E tests for global value inheritance
+[INFO] Using kubectl context: minikube
+[INFO] Test 1: Proxy configuration propagation
+[INFO] Installing chart with global.proxy...
+[INFO] Waiting for DaemonSet to be ready (timeout: 120s)...
+[INFO] DaemonSet ready: 1/1 pods
+[INFO] ✓ PASS: HTTP_PROXY environment variable set correctly
+[INFO] ✓ PASS: HTTPS_PROXY environment variable set correctly
+...
+========================================
+Test Summary
+========================================
+Tests run:    8
+Tests passed: 8
+Tests failed: 0
+========================================
+[INFO] All tests passed!
+```
+
+### Test Duration
+
+- **Per test**: ~30-60 seconds (Helm install + DaemonSet ready wait)
+- **Total suite**: ~5-8 minutes
+
+## Test Details
+
+### Test 1: Proxy Configuration Propagation
+
+**What it tests**: `global.proxy` → `HTTP_PROXY`/`HTTPS_PROXY` environment variables
+
+**Steps**:
+1. Install chart with `global.proxy: "http://test-proxy.example.com:3128"`
+2. Wait for DaemonSet to be ready
+3. Inspect pod spec for HTTP_PROXY and HTTPS_PROXY env vars
+4. Validate values match expected proxy URL
+
+**Why it matters**: Corporate environments require proxy configuration for outbound connections.
+
+---
+
+### Test 2: Proxy Override
+
+**What it tests**: Local `proxy` value takes precedence over `global.proxy`
+
+**Steps**:
+1. Install chart with both `global.proxy` and `proxy` set
+2. Verify HTTP_PROXY uses local `proxy` value (not global)
+
+**Why it matters**: Validates precedence model (local > global > default).
+
+---
+
+### Test 3: NodeSelector Propagation
+
+**What it tests**: `global.nodeSelector` applies to DaemonSet pods
+
+**Steps**:
+1. Label cluster node with `test-label=true`
+2. Install chart with `global.nodeSelector.test-label: "true"`
+3. Verify pod has nodeSelector field with correct label
+4. Cleanup: Remove node label
+
+**Why it matters**: Enables node targeting for dedicated monitoring nodes.
+
+---
+
+### Test 4: Tolerations Propagation
+
+**What it tests**: `global.tolerations` allows pods to tolerate node taints
+
+**Steps**:
+1. Taint cluster node with `test-taint=true:NoSchedule`
+2. Install chart with `global.tolerations` matching taint
+3. Verify pod schedules despite taint (has toleration in spec)
+4. Cleanup: Remove node taint
+
+**Why it matters**: Enables deployment on tainted nodes (e.g., monitoring-dedicated nodes).
+
+---
+
+### Test 5: Custom Registry Propagation
+
+**What it tests**: `global.images.registry` changes container image paths
+
+**Steps**:
+1. Install chart with `global.images.registry: "custom-registry.example.com"`
+2. Inspect pod spec for container image path
+3. Verify image path includes custom registry
+
+**Why it matters**: Critical for air-gapped environments and private registries.
+
+---
+
+### Test 6: ServiceAccount Annotations Propagation
+
+**What it tests**: `global.serviceAccount.annotations` propagate to ServiceAccount
+
+**Steps**:
+1. Install chart with `global.serviceAccount.annotations.eks.amazonaws.com/role-arn`
+2. Inspect ServiceAccount for annotation
+3. Verify annotation value matches expected
+
+**Why it matters**: Required for IAM roles (AWS IRSA, GCP Workload Identity, Azure Pod Identity).
+
+---
+
+### Test 7: VerboseLog Propagation
+
+**What it tests**: `global.verboseLog: true` → `LOG_LEVEL=debug`
+
+**Steps**:
+1. Install chart with `global.verboseLog: true`
+2. Inspect pod spec for LOG_LEVEL env var
+3. Verify value is "debug" (not default "info")
+
+**Why it matters**: Enables debug logging for troubleshooting.
+
+---
+
+### Test 8: HostNetwork Propagation
+
+**What it tests**: `global.hostNetwork: true` enables host network mode
+
+**Steps**:
+1. Install chart with `global.hostNetwork: true`
+2. Inspect pod spec for hostNetwork field
+3. Verify field is true
+
+**Why it matters**: Required for certain networking configurations (e.g., UDP log forwarding).
+
+---
+
+## Troubleshooting
+
+### Test Failures
+
+#### "DaemonSet did not become ready within 120s"
+
+**Cause**: Image not available, insufficient resources, or scheduling constraints preventing pod from running.
+
+**Solutions**:
+1. Verify image is loaded into Minikube: `minikube image ls | grep newrelic-fluentbit`
+2. Check pod events: `kubectl describe pod <pod-name>`
+3. Check pod logs: `kubectl logs <pod-name>`
+4. Verify cluster has resources: `kubectl top nodes`
+
+#### "NodeSelector test failing"
+
+**Cause**: Node label not applied or multiple nodes present.
+
+**Solutions**:
+1. Verify node has label: `kubectl get nodes --show-labels`
+2. If multiple nodes, ensure label is on all nodes or adjust test
+
+#### "Tolerations test failing"
+
+**Cause**: Taint not applied or pod doesn't have toleration.
+
+**Solutions**:
+1. Verify node has taint: `kubectl describe node <node-name> | grep Taints`
+2. Check pod tolerations: `kubectl get pod <pod-name> -o yaml | grep -A 10 tolerations`
+
+### Cleanup Issues
+
+If tests fail midway, you may need to manually clean up:
+
+```bash
+# Delete Helm release
+helm delete nr-logging-e2e --namespace default
+
+# Remove node labels
+kubectl label node --all test-label-
+
+# Remove node taints
+kubectl taint node --all test-taint-
+
+# Delete test pods
+kubectl delete pod test-pod-for-exec --namespace default
+```
+
+## Integration with CI/CD
+
+These tests are designed to run in GitHub Actions workflows. See `.github/workflows/functional-e2e.yaml` for the workflow configuration.
+
+### Workflow Triggers
+
+- **On PR**: Run against all supported Kubernetes versions
+- **On Push to main**: Run regression tests
+- **Manual**: Via workflow_dispatch
+
+### Kubernetes Version Matrix
+
+Tests run against:
+- v1.34.0
+- v1.33.0
+- v1.32.0
+- v1.31.0
+- v1.30.0
+
+## Limitations
+
+### What These Tests Don't Cover
+
+1. **Telemetry Validation**: Tests don't verify that logs reach New Relic (use `newrelic-integration-e2e-action` for that)
+2. **Fluent Bit Functionality**: Tests don't verify log parsing, filtering, or forwarding logic
+3. **Multi-Node Scheduling**: Tests assume single-node cluster (Minikube)
+4. **Windows DaemonSet**: Tests only cover Linux DaemonSet (Windows requires Windows node)
+5. **Performance**: No load testing or resource usage validation
+
+### Known Issues
+
+- **Image Loading**: Tests require pre-built image loaded into Minikube (not pulled from registry)
+- **Single Node**: Tests may not catch multi-node scheduling issues
+- **No Proxy Server**: Tests validate proxy env vars but don't test actual proxy connectivity
+
+## Future Enhancements
+
+1. **Multi-Chart Tests**: Validate global values across multiple charts in nri-bundle
+2. **Real Proxy Test**: Deploy Squid proxy in Minikube and validate connectivity
+3. **Air-Gapped Registry Test**: Deploy local registry and validate image pull
+4. **Windows Support**: Add Windows node to Minikube and test Windows DaemonSet
+5. **Fargate Exclusion Test**: Validate that affinity rules exclude Fargate nodes
+6. **Priority Class Test**: Validate `global.priorityClassName` propagation
+7. **Affinity Test**: Validate `global.affinity` propagation
+
+## Related Documentation
+
+- [Helm Unit Tests](../README.md): Template-level tests using helm-unittest
+- [E2E Telemetry Tests](../../../../e2e/README.md): Telemetry validation tests (if available)
+- [Global Values Specification](../../../../CLAUDE.md): Complete list of 27 global values
+
+## Support
+
+For issues or questions:
+- **GitHub Issues**: [newrelic/helm-charts/issues](https://github.com/newrelic/helm-charts/issues)
+- **Pull Requests**: Contributions welcome!
+- **Documentation**: [New Relic Kubernetes Integration Docs](https://docs.newrelic.com/docs/kubernetes-pixie/kubernetes-integration/get-started/introduction-kubernetes-integration/)