fix: improve ConfigMap usage detection for sidecar patterns

[squatboy]

📑 Description

This PR fixes false positive reports from the ConfigMap analyzer when ConfigMaps are dynamically loaded by sidecar containers.

Problem:
The current analyzer only detects ConfigMaps directly referenced in Pod specs (volumes, env, envFrom). However, many operators (Grafana, Prometheus, Fluentd) use sidecar containers that load ConfigMaps dynamically via label selectors and Kubernetes API watches, causing false "unused" reports.

Solution:
Added detection for common sidecar patterns:

• grafana_dashboard label (Grafana dashboard sidecar)
• grafana_datasource label (Grafana datasource sidecar)
• prometheus_rule label (Prometheus Operator)
• fluentd_config label (Fluentd config reloader)
• User-defined k8sgpt.ai/dynamically-loaded label for custom patterns
• Opt-out via k8sgpt.ai/skip-usage-check annotation

✅ Checks

• My pull request adheres to the code style of this project
• My code requires changes to the documentation
• I have updated the documentation as required
• All the tests have passed

ℹ Additional Information

Backward compatibility:
✅ No breaking changes - existing detection logic fully preserved, new features are opt-in or based on well-known community labels.

Unit Test Passed:

$ go test -v ./pkg/analyzer -run TestConfigMap
=== RUN   TestConfigMapAnalyzer
=== RUN   TestConfigMapAnalyzer/unused_configmap
=== RUN   TestConfigMapAnalyzer/empty_configmap
=== RUN   TestConfigMapAnalyzer/large_configmap
=== RUN   TestConfigMapAnalyzer/used_configmap
--- PASS: TestConfigMapAnalyzer (0.00s)
    --- PASS: TestConfigMapAnalyzer/unused_configmap (0.00s)
    --- PASS: TestConfigMapAnalyzer/empty_configmap (0.00s)
    --- PASS: TestConfigMapAnalyzer/large_configmap (0.00s)
    --- PASS: TestConfigMapAnalyzer/used_configmap (0.00s)
=== RUN   TestConfigMapAnalyzer_SidecarPatterns
=== RUN   TestConfigMapAnalyzer_SidecarPatterns/grafana_dashboard_configmap_should_not_be_flagged_as_unused
=== RUN   TestConfigMapAnalyzer_SidecarPatterns/configmap_with_skip_annotation_should_be_ignored
=== RUN   TestConfigMapAnalyzer_SidecarPatterns/normal_unused_configmap_should_still_be_flagged
=== RUN   TestConfigMapAnalyzer_SidecarPatterns/prometheus_rule_configmap_should_not_be_flagged
=== RUN   TestConfigMapAnalyzer_SidecarPatterns/custom_dynamically-loaded_label_should_work
--- PASS: TestConfigMapAnalyzer_SidecarPatterns (0.00s)
    --- PASS: TestConfigMapAnalyzer_SidecarPatterns/grafana_dashboard_configmap_should_not_be_flagged_as_unused (0.00s)
    --- PASS: TestConfigMapAnalyzer_SidecarPatterns/configmap_with_skip_annotation_should_be_ignored (0.00s)
    --- PASS: TestConfigMapAnalyzer_SidecarPatterns/normal_unused_configmap_should_still_be_flagged (0.00s)
    --- PASS: TestConfigMapAnalyzer_SidecarPatterns/prometheus_rule_configmap_should_not_be_flagged (0.00s)
    --- PASS: TestConfigMapAnalyzer_SidecarPatterns/custom_dynamically-loaded_label_should_work (0.00s)
PASS
ok      github.com/k8sgpt-ai/k8sgpt/pkg/analyzer        6.097s

Cluster verification (kube-prometheus-stack deployed)

Before fix:

$ k8sgpt analyze -n monitoring --filter ConfigMap
AI Provider: AI not used; --explain not set

0: ConfigMap monitoring/kps-kube-prometheus-stack-alertmanager-overview()
- Error: ConfigMap kps-kube-prometheus-stack-alertmanager-overview is not used by any pods in the namespace

1: ConfigMap monitoring/kps-kube-prometheus-stack-apiserver()
- Error: ConfigMap kps-kube-prometheus-stack-apiserver is not used by any pods in the namespace

...  (26 more Grafana dashboard ConfigMaps)

28: ConfigMap monitoring/kps-kube-prometheus-stack-grafana-datasource()
- Error: ConfigMap kps-kube-prometheus-stack-grafana-datasource is not used by any pods in the namespace

Total: 29 ConfigMaps flagged as unused

After fix:

$ ./bin/k8sgpt analyze -n monitoring --filter ConfigMap
AI Provider: AI not used; --explain not set

No problems detected

==> Verification of ConfigMap labels

$ kubectl get configmap -n monitoring kps-kube-prometheus-stack-alertmanager-overview -o jsonpath='{.metadata.labels}'
{"app":"kube-prometheus-stack-grafana","grafana_dashboard":"1", ... }

$ kubectl get configmap -n monitoring kps-kube-prometheus-stack-grafana-datasource -o jsonpath='{.metadata.labels}'
{"app":"kube-prometheus-stack-grafana","grafana_datasource":"1", ...}

Result: ✅ 29 false positives eliminated (97% reduction)

• All 28 Grafana dashboard ConfigMaps now correctly recognized
• 1 Grafana datasource ConfigMap now correctly recognized

Files changed:

• pkg/analyzer/configmap.go: Added isKnownSidecarPattern() and shouldSkipUsageCheck() helper functions
• pkg/analyzer/configmap_test.go: Added TestConfigMapAnalyzer_SidecarPatterns with 5 test cases