Add gcp-compute-instancegroup-health CodeBundle

[rw-codebundle-agent]

Closes #153

Adds the gcp-compute-instancegroup-health CodeBundle to rw-cli-codecollection.

Summary

Monitors the health of GCP Compute Engine instance groups (managed and
unmanaged) at the group scope — one SLX per instance group rather than one
per member VM, as required by the parent issue.

Tasks

• Discover GCP Instance Groups and Configurations (discover_instance_groups.sh)
• Check Instance Group Member Health (check_group_member_health.sh)
• Check Instance Group Autoscaling and Capacity (check_autoscaling.sh)
• Check Instance Group OS Patch Compliance (check_group_patch_status.sh)
• Check Instance Group Utilization (check_group_utilization.sh)
• Generate Instance Group Health Summary (generate_group_summary.sh)

SLI

In-repo sli.robot produces a 0-1 health score averaged across four dimensions:
member health, autoscaling, patch compliance, and utilization. Each dimension is
pushed as a sub-metric, with the aggregate pushed as the primary metric.

Notes / Deviations from Design Spec

• Generation rules use the CloudQuery table names (gcp_compute_instance_groups,
  gcp_compute_instance_group_managers) and qualifiers: ["resource"] to
  generate one SLX per instance group.
• INSTANCE_GROUP_NAME is auto-set from the matched resource in the templates.
• Auth uses the standard gcp-auth.yaml include with gcp_credentials secret.

Scoring

python -m scorer.score → 104/104 (PASS, threshold 70).